Peng et al. (2025) Spatiotemporal Reconstruction of Annual Glacier Mass Balance in Central Asia (2000–2020) Using Machine Learning

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

Identification

• Journal: Journal of Geophysical Research Atmospheres
• Year: 2025
• Date: 2025-09-12
• Authors: Yanfei Peng, Tobias Bolch, Qiangqiang Yuan, Francesca Baldacchino, Qianqian Yang
• DOI: 10.1029/2024jd043191

Research Groups

Not available in the abstract.

Short Summary

This study reconstructs annual glacier-wide mass balance for glaciers in the Tien Shan and Pamir from 2000 to 2020 using machine learning, revealing an average mass loss of -0.39 meters water equivalent per year with significant spatiotemporal variability and accelerated loss for smaller glaciers.

Objective

• To reconstruct annual glacier-wide mass balance from 2000 to 2020 for glaciers larger than 0.1 square kilometers across the Tien Shan and Pamir using machine learning techniques.

Study Configuration

• Spatial Scale: Glaciers larger than 0.1 square kilometers across the Tien Shan and Pamir regions.
• Temporal Scale: 2000 to 2020 (21 years).

Methodology and Data

• Models used: Machine learning (ML) techniques, including five ensemble ML models and a deep neural network; XGBoost was selected for the final reconstruction due to its superior performance.
• Data sources: Meteorological data from the ERA5-Land dataset and topographic features.

Main Results

• An average glacier-wide mass loss of -0.39 meters water equivalent per year was observed for the studied period.
• The highest mass losses occurred in the Djungar Alatau region, averaging -0.68 meters water equivalent per year.
• The lowest mass losses were found in the eastern Pamir, averaging -0.10 meters water equivalent per year.
• Small glaciers (area less than 1 square kilometer) experienced more rapid mass loss compared to larger glaciers.
• The temporal evolution of glacier mass balance showed an average acceleration in mass loss, with notable spatiotemporal variability.
• Glacier elevation and geographic location were identified as the dominant factors influencing mass balance, followed by July and August temperatures.

Contributions

• Advances the application of machine learning methods in glaciology.
• Enhances the understanding of regional glacier mass balance in High-mountain Asia, particularly for the Tien Shan and Pamir regions, over a significant spatial and temporal extent.

Funding

Not available in the abstract.

Citation

@article{Peng2025Spatiotemporal,
  author = {Peng, Yanfei and Bolch, Tobias and Yuan, Qiangqiang and Baldacchino, Francesca and Yang, Qianqian},
  title = {Spatiotemporal Reconstruction of Annual Glacier Mass Balance in Central Asia (2000–2020) Using Machine Learning},
  journal = {Journal of Geophysical Research Atmospheres},
  year = {2025},
  doi = {10.1029/2024jd043191},
  url = {https://doi.org/10.1029/2024jd043191}
}