Wang et al. (2026) A Data-Driven Framework for Characterizing Nonlinear Responses of River Ice Growth and Decay in the Shisifenzi Reach of the Upper Yellow River

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

Identification

• Journal: Water
• Year: 2026
• Date: 2026-09-16
• Authors: Jingwen Wang, Heli Yu, Yusu Yue, Yu Deng, Lianjun Zhao, Shasha Han, Ziyang Wei, Ming Luo
• DOI: 10.3390/w18182315

Research Groups

• Institute of Hydrology and Water Resources Engineering, Tongji University
• Key Laboratory of Water Cycle and Water Security in Yangtze River Basin, Ministry of Education

Short Summary

This study develops a data-driven framework using Random Forest with SHAP and ALE to investigate the nonlinear responses of river ice growth and decay to hydro-thermal features in the Shisifenzi Reach of the Upper Yellow River. The main finding is that the 30-day cumulative freezing degree-hours is the dominant driver, with a threshold near 4000 °C·h separating slow growth from rapid thickening.

Objective

• Investigate the nonlinear responses of river ice growth and decay to hydro-thermal features in the Shisifenzi Reach of the Upper Yellow River

Study Configuration

• Spatial Scale: Localized at the Shisifenzi Reach, Upper Yellow River
• Temporal Scale: Six winter seasons (2020–2026)

Methodology and Data

• Models used: Random Forest with SHAP and Accumulated Local Effects (ALE)
• Data sources: Monitoring data from six winter seasons (2020–2026) at the Shisifenzi Reach

Main Results

• The 30-day cumulative freezing degree-hours is the dominant driver, with a threshold near 4000 °C·h separating slow growth from rapid thickening.
• Water level elevation shows an obvious nonlinear link: ice is thickest when it stays in the 988.5–989 m range; above this, the ALE effect drops due to slight flow velocity disturbance and high water level coinciding with warmer temperature.
• Short-term thermal features show little effects, while other thermal variables and feature interactions also show saturation.

Contributions

• The study provides a data-driven framework for interpretable, site-specific information useful for river ice management.
• The findings highlight the importance of considering nonlinear responses to hydro-thermal features in river ice research.

Funding

• This research was funded by the National Natural Science Foundation of China (Grant No. 51879224) and the Ministry of Education's Key Laboratory of Water Cycle and Water Security in Yangtze River Basin (Grant No. KF202101).

Citation

@article{Wang2026DataDriven,
  author = {Wang, Jingwen and Yu, Heli and Yue, Yusu and Deng, Yu and Zhao, Lianjun and Han, Shasha and Wei, Ziyang and Luo, Ming},
  title = {A Data-Driven Framework for Characterizing Nonlinear Responses of River Ice Growth and Decay in the Shisifenzi Reach of the Upper Yellow River},
  journal = {Water},
  year = {2026},
  doi = {10.3390/w18182315},
  url = {https://doi.org/10.3390/w18182315}
}