An et al. (2026) Signal-domain guided deep learning for gap-filling of XCO and XCH 4 : a masked spatio-temporal fusion of TROPOMI and GEOS-Chem (2019–2023)

Identification

• Journal: Earth system science data
• Year: 2026
• Date: 2026-09-16
• Authors: Chengkun An, Yuan Tian, Zhiwei Li, Qiaoyu Jiang, Peize Lin, Bowen Chang, Jingkai Xue, Youwen Sun
• DOI: 10.5194/essd-18-6859-2026

Research Groups

• Institutes of Physical Science and Information Technology, Anhui University, Hefei 230601, China
• National Key Laboratory of Opto-Electronic Information Acquisition and Protection Technology, Institutes of Physical Science and Information Technology, Anhui University, Hefei, Anhui, China
• School of Environmental Science and Optoelectronic Technology, University of Science and Technology of China, Hefei 230026, China
• Key Laboratory of Environmental Optics and Technology, Anhui Institute of Optics and Fine Mechanics, HFIPS, Chinese Academy of Sciences, Hefei 230031, China

Short Summary

This study proposes a novel signal-domain guided spatio-temporal fusion framework to generate daily global and regional continuous XCO and XCH4 products (2019–2023) at high resolution. The method effectively leverages complementary information from chemical transport modeling and frequency-domain representations.

Objective

• To develop an efficient and interpretable solution for large-scale trace gas monitoring by integrating physical modeling, signal-domain reconstruction, and deep learning-based residual correction.
• To generate daily global and regional continuous XCO and XCH4 products (2019–2023) at high resolution using a two-stage fusion framework.

Study Configuration

• Spatial Scale: Global and regional scales with spatial resolutions of 0.25° globally and 0.05° over China.
• Temporal Scale: Daily temporal scale from 2019 to 2023.

Methodology and Data

• Models used:
  • TROPOMI Level-2 data products of column-averaged CO (XCO) and CH4 (XCH4).
  • GEOS-Chem chemical transport model.
• Data sources:
  • Satellite observations from TROPOMI.
  • Chemical transport modeling simulations from GEOS-Chem.

Main Results

• The proposed framework generates daily global and regional continuous XCO and XCH4 products (2019–2023) at high resolution.
• The fused outputs outperform GEOS-Chem simulations alone and maintain accuracy comparable to or better than original TROPOMI retrievals in cloud-covered regions.
• The method effectively leverages complementary information from chemical transport modeling and frequency-domain representations.

Contributions

• This study proposes a novel signal-domain guided spatio-temporal fusion framework for large-scale trace gas monitoring.
• The method provides an efficient and interpretable solution for generating daily global and regional continuous XCO and XCH4 products (2019–2023) at high resolution.

Funding

• National Key Laboratory of Opto-Electronic Information Acquisition and Protection Technology, Institutes of Physical Science and Information Technology, Anhui University.
• School of Environmental Science and Optoelectronic Technology, University of Science and Technology of China.

Citation

@article{An2026Signaldomain,
  author = {An, Chengkun and Tian, Yuan and Li, Zhiwei and Jiang, Qiaoyu and Lin, Peize and Chang, Bowen and Xue, Jingkai and Sun, Youwen},
  title = {Signal-domain guided deep learning for gap-filling of XCO and XCH 4 : a masked spatio-temporal fusion of TROPOMI and GEOS-Chem (2019–2023)},
  journal = {Earth system science data},
  year = {2026},
  doi = {10.5194/essd-18-6859-2026},
  url = {https://doi.org/10.5194/essd-18-6859-2026}
}