Chen et al. (2026) Democratizing planetary-scale analysis: an ultra-lightweight Earth embedding database for accurate and flexible global land monitoring

Identification

• Journal: Earth system science data
• Year: 2026
• Date: 2026-07-24
• Authors: Shuang Chen, Jie Wang, Shuai Yuan, Jiayang Li, Yu Xia, yuanhong Liao, Junbo Wei, Jincheng Yuan, Xiaoqing Xu, Xiaolin Zhu, Peng Zhu, Hailong Zhang, Y Zhou, Haohuan Fu, Huabing Huang, Бин Чэн, Fan Dai, Peng Gong
• DOI: 10.5194/essd-18-5375-2026

Research Groups

• Department of Geography, The University of Hong Kong, Hong Kong, China
• Pengcheng Laboratory, Shenzhen 518000, China
• Ministry of Education Key Laboratory for Earth System Modeling, Department of Earth System Science, Tsinghua University, Beijing 100084, China
• Institute for Climate and Carbon Neutrality, The University of Hong Kong, Hong Kong, China
• School of Geospatial Engineering and Science, Sun Yat-sen University, Guangzhou 510275, China

Short Summary

This paper presents the Embedded Seamless Data (ESD), an ultra-lightweight Earth embedding database for accurate and flexible global land monitoring. The ESD dataset is designed to overcome the barriers in storage and computational resources required for planetary-scale analysis.

Objective

• To develop a global-scale 30m Earth embedding dataset that can be used for accurate and efficient environmental monitoring.
• To address the challenges of storing and processing large-scale remote sensing data.

Study Configuration

• Spatial Scale: Global, with a spatial resolution of 30m.
• Temporal Scale: 25-year period from 2000 to 2024.

Methodology and Data

• Models used: ESDNet architecture and Finite Scalar Quantization (FSQ) module.
• Data sources: Landsat series (5, 7, 8, and 9), MODIS Terra, NASADEM Global Digital Elevation Model, GAIA dataset, ESA WorldCover 2021, GLAD global crop extent dynamics, and GLAD global monthly surface water dynamics.

Main Results

• The ESD dataset achieves a transformative ∼340-fold reduction in data volume compared to raw daily archives.
• The entire global land surface for a single year is encapsulated within approximately 2.4TB.
• Rigorous validation demonstrates that ESD maintains high reconstructive fidelity to the original reflectance values across the spectral dimension, achieving a Mean Absolute Error (MAE) of 0.0130 and a Root Mean Square Error (RMSE) of 0.0179.

Contributions

• The ESD dataset provides a versatile foundation for democratizing planetary-scale Earth system research and advancing next-generation geospatial artificial intelligence.
• The dataset offers several transformative advantages over existing products, including high compression ratios, extended temporal coverage, and the explicit preservation of temporal structures.

Funding

• This work was supported by the National Natural Science Foundation of China (Grant Nos. 42122015, 41790434) and the Research Grants Council of Hong Kong (Grant No. HKU17300820).

Citation

@article{Chen2026Democratizing,
  author = {Chen, Shuang and Wang, Jie and Yuan, Shuai and Li, Jiayang and Xia, Yu and Liao, yuanhong and Wei, Junbo and Yuan, Jincheng and Xu, Xiaoqing and Zhu, Xiaolin and Zhu, Peng and Zhang, Hailong and Zhou, Y and Fu, Haohuan and Huang, Huabing and Чэн, Бин and Dai, Fan and Gong, Peng},
  title = {Democratizing planetary-scale analysis: an ultra-lightweight Earth embedding database for accurate and flexible global land monitoring},
  journal = {Earth system science data},
  year = {2026},
  doi = {10.5194/essd-18-5375-2026},
  url = {https://doi.org/10.5194/essd-18-5375-2026}
}