Feng et al. (2026) Research on Hydrological Prediction Based on Improved Successive Variational Mode Decomposition and Dual-Attention Temporal Convolutional Network

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

Identification

• Journal: Atmosphere
• Year: 2026
• Date: 2026-09-24
• Authors: Wenwen Feng, Shilei Zhang, Xiaohui Lei, Xiaodong Xu, Yang W. Shao, Chao Wang, Zhi Ye, Chu Zhang
• DOI: 10.3390/atmos17100927

Research Groups

• Institute of Hydrology and Water Resources, Sichuan University
• Department of Hydraulic Engineering, Tongji University

Short Summary

This study proposes a novel hybrid model combining Improved Exponential–Trigonometric Optimization (IETO), Singular Value Decomposition (SVMD), and dual-attention temporal convolutional network (DATCN) for runoff prediction in the middle reaches of the Jinsha River. The proposed IETO-SVMD-DATCN model achieves improved forecasting performance compared to its individual components.

Objective

• Investigate the effectiveness of a hybrid model combining IETO, SVMD, and DATCN for runoff prediction in the middle reaches of the Jinsha River.

Study Configuration

• Spatial Scale: Middle reaches of the Jinsha River
• Temporal Scale: Hourly to daily time scales

Methodology and Data

• Models used:
  • Improved Exponential–Trigonometric Optimization (IETO)
  • Singular Value Decomposition (SVMD)
  • Dual-attention temporal convolutional network (DATCN)
• Data sources:
  • Precipitation and runoff data from five monitoring stations

Main Results

• The proposed IETO-SVMD-DATCN model achieved an NSE of 0.9826, outperforming individual components.
• Lower RMSE, MAE, and MAPE values were obtained with the proposed model.

Contributions

• This study contributes to improved runoff prediction capabilities in the middle reaches of the Jinsha River by proposing a novel hybrid model combining IETO, SVMD, and DATCN.
• The results demonstrate enhanced forecasting performance compared to existing models.

Funding

• National Natural Science Foundation of China (Grant No. 51879215)
• Sichuan Provincial Key Research and Development Program (Grant No. 2021YJ0113)

Citation

@article{Feng2026Research,
  author = {Feng, Wenwen and Zhang, Shilei and Lei, Xiaohui and Xu, Xiaodong and Shao, Yang W. and Wang, Chao and Ye, Zhi and Zhang, Chu},
  title = {Research on Hydrological Prediction Based on Improved Successive Variational Mode Decomposition and Dual-Attention Temporal Convolutional Network},
  journal = {Atmosphere},
  year = {2026},
  doi = {10.3390/atmos17100927},
  url = {https://doi.org/10.3390/atmos17100927}
}