Jin et al. (2026) TPHH: a long-term (1901–2023) high-resolution (1∕30°) near-surface humidity dataset for the Tibetan Plateau generated via spatial downscaling based on hybrid-structure deep learning

Identification

• Journal: Earth system science data
• Year: 2026
• Date: 2026-09-14
• Authors: Zheng Zhong Jin, Zezhou Chen, Qinglong You, Jintao Zhang, Huan Hu, Ping Chen, Xiang Liu, Zipeng Wang, Kai Wang, Shiguo Lian, Shichang Kang
• DOI: 10.5194/essd-18-6763-2026

Research Groups

• College of Geography and Planning, Chengdu University of Technology
• Data Science & Artificial Intelligence Research Institute, China Unicom
• Department of Atmospheric and Oceanic Sciences & Institute of Atmospheric Sciences, Fudan University
• Yunnan Key Laboratory of Plateau Geographical Process and Environmental Changes, Faculty of Geography, Yunnan Normal University
• Institute of Mountain Hazards and Environment, Chinese Academy of Sciences

Short Summary

This study generates a long-term high-resolution near-surface humidity dataset (TPHH) for the Tibetan Plateau using a hybrid-structure deep learning framework. The dataset covers 1901–2023 at a spatial resolution of 1/30° × 1/30°.

Objective

• To develop a high-resolution climate baseline including both temperature and humidity from the early 20th century onwards to characterize changes across the highly rugged terrain of the Tibetan Plateau.
• To reconstruct historical high-resolution climate fields over the TP using a spatial downscaling framework that maps centennial-scale CRU signals onto high-resolution grids.

Study Configuration

• Spatial Scale: The study focuses on the Tibetan Plateau, with a spatial resolution of 1/30° × 1/30° (approximately 3 km × 3 km).
• Temporal Scale: The dataset covers the period from 1901 to 2023, with a monthly temporal resolution.

Methodology and Data

• Models used: FourCastNet, a hybrid-structure deep learning model combining Vision Transformers (ViT) and Adaptive Fourier Neural Operators (AFNO).
• Data sources: Climatic Research Unit Time-Series v4.08 (CRU TS v4.08), Tibetan Plateau Multi-source Meteorological Forcing Dataset (TPMFD).

Main Results

• The TPHH dataset provides reconstructed monthly 2 m temperature, 2 m specific humidity, and surface pressure at 1/30° × 1/30° resolution for 1901–2023.
• Comparisons with gridded products and station-based consistency evaluations indicate that TPHH reproduces several large-scale and station-observed statistical features.

Contributions

• The study contributes to the development of a long-term high-resolution climate baseline for the Tibetan Plateau, which is essential for characterizing changes across its highly rugged terrain.
• The use of FourCastNet as a downscaling model demonstrates its ability to efficiently learn multi-scale spatial features and total-field dependencies, making it suitable for climate data downscaling tasks.

Funding

• This research was funded by [insert funding projects, programs, and reference codes].

Citation

@article{Jin2026TPHH,
  author = {Jin, Zheng Zhong and Chen, Zezhou and You, Qinglong and Liu, Zhaoxiang and Zhang, Jintao and Hu, Huan and Chen, Ping and Liu, Xiang and Wang, Zipeng and Wang, Kai and Lian, Shiguo and Kang, Shichang},
  title = {TPHH: a long-term (1901–2023) high-resolution (1∕30°) near-surface humidity dataset for the Tibetan Plateau generated via spatial downscaling based on hybrid-structure deep learning},
  journal = {Earth system science data},
  year = {2026},
  doi = {10.5194/essd-18-6763-2026},
  url = {https://doi.org/10.5194/essd-18-6763-2026}
}