Wang et al. (2026) PhyGroup-UNet: Dynamic Physically Correlated Channel Grouping in a Lightweight UNet for Efficient Precipitation Nowcasting

Identification

• Journal: IEEE Transactions on Geoscience and Remote Sensing
• Year: 2026
• Date: 2026-01-01
• Authors: Chuang Wang, Bo Sang, Qiang He, Shumin Han
• DOI: 10.1109/tgrs.2026.3667511

Research Groups

[Not specified in the provided text.]

Short Summary

This paper introduces PhyGroup-UNet, a lightweight UNet architecture that employs dynamic physically correlated channel grouping to enhance the efficiency of precipitation nowcasting.

Objective

• To develop an efficient deep learning model (PhyGroup-UNet) for precipitation nowcasting by integrating dynamic physically correlated channel grouping within a lightweight UNet framework.

Study Configuration

• Spatial Scale: [Not specified in the provided text.]
• Temporal Scale: Nowcasting (implying short-term prediction, typically minutes to a few hours).

Methodology and Data

• Models used: PhyGroup-UNet (a lightweight UNet variant with dynamic physically correlated channel grouping).
• Data sources: [Not specified in the provided text, but typically involves radar or satellite precipitation data for nowcasting.]

Main Results

[Not specified in the provided text.]

Contributions

• Introduction of "Dynamic Physically Correlated Channel Grouping" as a novel mechanism.
• Development of "PhyGroup-UNet," a lightweight UNet architecture specifically designed for efficient precipitation nowcasting.

Funding

[Not specified in the provided text.]

Citation

@article{Wang2026PhyGroupUNet,
  author = {Wang, Chuang and Sang, Bo and He, Qiang and Han, Shumin},
  title = {PhyGroup-UNet: Dynamic Physically Correlated Channel Grouping in a Lightweight UNet for Efficient Precipitation Nowcasting},
  journal = {IEEE Transactions on Geoscience and Remote Sensing},
  year = {2026},
  doi = {10.1109/tgrs.2026.3667511},
  url = {https://doi.org/10.1109/tgrs.2026.3667511}
}