Ajao et al. (2026) Diagnosing the Roles of Meteorological Forcing, Storage Memory, and Network Connectivity in Cold-Region Streamflow Prediction

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

Identification

• Journal: Water Resources Management
• Year: 2026
• Date: 2026-08-27
• Authors: Adewale Ajao, Yeo Howe Lim
• DOI: 10.1007/s11269-026-04869-9

Research Groups

• Department of Hydrology, University of Alaska Fairbanks
• Institute for Arctic and Alpine Research (INSTAAR), University of Colorado Boulder

Short Summary

This study proposes a six-tier diagnostic framework to quantify the contributions of meteorological forcing, storage memory, and network connectivity to streamflow prediction in cold-region watersheds. The results show that graph-based models provide greater physical realism and transferability despite slightly lower predictive accuracy than autoregressive approaches.

Objective

• Investigate the relative contributions of meteorological forcing, hydrologic memory, and spatial watershed connectivity to streamflow prediction in cold-region watersheds

Study Configuration

• Spatial Scale: The study was conducted in the Goose River watershed, a cold-region watershed located in Alaska.
• Temporal Scale: The study analyzed daily streamflow data from 2000 to 2019.

Methodology and Data

• Models used:
  • Temporal deep learning models
  • Static machine-learning models
  • Graph neural networks (GNNs)
  • Spatio-temporal GNNs
• Data sources: Daily streamflow data from the Goose River watershed, meteorological forcing data from a nearby weather station

Main Results

• Meteorological forcing dominated short-term discharge prediction.
• Incorporating antecedent discharge memory produced the highest numerical performance (NSE = 0.989).
• The spatio-temporal GNN achieved the most physically consistent representation by integrating dynamic forcing with upstream–downstream connectivity.

Contributions

• This study provides a six-tier diagnostic framework to quantify the contributions of meteorological forcing, storage memory, and network connectivity to streamflow prediction in cold-region watersheds.
• The results demonstrate a trade-off between statistical accuracy and hydrologic interpretability, highlighting the importance of considering physical realism and transferability when selecting models for streamflow prediction.

Funding

• This research was supported by the National Science Foundation (Award #1835841) and the University of Alaska Fairbanks' Research Initiative.

Citation

@article{Ajao2026Diagnosing,
  author = {Ajao, Adewale and Lim, Yeo Howe},
  title = {Diagnosing the Roles of Meteorological Forcing, Storage Memory, and Network Connectivity in Cold-Region Streamflow Prediction},
  journal = {Water Resources Management},
  year = {2026},
  doi = {10.1007/s11269-026-04869-9},
  url = {https://doi.org/10.1007/s11269-026-04869-9}
}