Researchers at the Northwest Institute of Eco-Environment and Resources (NIEER), Chinese Academy of Sciences, working with Nanjing University of Information Science and Technology and other institutions, have advanced fine-scale mapping of permafrost occurrence and active-layer thickness in the Wenquan area on the northeastern Qinghai–Tibet Plateau.

Combining ground-penetrating radar (GPR), field investigations, remote sensing and machine learning, the team developed 30-m permafrost distribution maps using 1,037 samples from multiple sources.

The predicted lower elevation limit of permafrost differed by more than 200 m between slope aspects, indicating substantial variation associated with local terrain and thermal conditions. In a related study, the researchers used 810 radar-derived samples from 2009 to reconstruct late-thaw active-layer thickness at 30-m resolution.

The model achieved an R² of 0.81 and a root mean square error of 0.27 m in held-out sample validation. Active-layer thickness generally increased from swamp meadows to desert steppes, with substantial variation within each landscape type.

Further analysis identified soil properties, snow-cover duration and thermal conditions as important predictors, with relationships that varied across landscapes.

Published in Remote Sensing and CATENA, the studies demonstrate the potential of combining geophysical observations with machine learning to map mountain permafrost. They also provide spatial information for permafrost monitoring, model evaluation and cold-region environmental research.

https://doi.org/10.3390/rs17122015

https://doi.org/10.1016/j.catena.2026.110744