Permafrost, a key component of the global cryosphere, has a vast distribution and contains substantial underground ice and carbon reservoirs. Its degradation poses severe threats to infrastructure safety, releases large amounts of long-term sequestered greenhouse gases, and affects precious freshwater resources. However, due to its subsurface location, the scientific community has lacked effective remote sensing methods to accurately monitor changes in its hydrological and thermal state. Traditional ground observation methods, while precise, are difficult to scale for large-area monitoring. Synthetic Aperture Radar Interferometry (InSAR) technology, by capturing subtle surface deformations, indirectly reflects the hydrological and thermal dynamics of permafrost, making it a critical tool for current permafrost monitoring. However, conventional InSAR time-series models have not sufficiently considered the unique and significant seasonal deformation characteristics of permafrost areas, a limitation that severely restricts monitoring accuracy and reliability. Addressing this technological bottleneck is crucial for accurately assessing permafrost changes and their environmental impact.
Recently, a research team led by Professor Mu Cuicui at the School of Earth and Environmental Sciences, Lanzhou University, published significant findings in the renowned journal ISPRS Journal of Photogrammetry and Remote Sensing (Impact Factor: 10.6). The paper, titled Time-Series Models for Ground Subsidence and Heave Over Permafrost in InSAR Processing: A Comprehensive Assessment and New Improvement, is now available for open access (Link:https://doi.org/10.1016/j.isprsjprs.2025.02.019). The study introduces innovative improvements to time-series models, addressing the unique seasonal surface deformation characteristics of permafrost areas. These advancements not only optimize the expression of surface deformation caused by freeze-thaw cycles but also introduce a novel physical-mechanism-based automatic reference point selection algorithm, significantly enhancing the precision and reliability of InSAR monitoring. These breakthroughs provide essential technical support and methodological advancements for surface deformation monitoring in permafrost regions.
The research team optimized InSAR signal processing capabilities, improving both a mathematical and a physical model. Specifically, they introduced a semi-annual cycle component into the sinusoidal function model, allowing it to capture seasonal deformation characteristics and the "fully frozen winter phase," thus significantly improving the model’s ability to represent freeze-thaw cycle deformation in permafrost areas. This mathematical model, due to its simplicity and efficiency, is highly recommended to replace the commonly used linear or quadratic models in InSAR processing software, reducing the loss of surface deformation signals in permafrost areas. Additionally, the physical model based on the Stefan equation was optimized, deepening the theoretical link between freeze-thaw cycles and surface deformation, and creating seamless stitching methods for two freeze-thaw cycles, which effectively improved the simulation accuracy of surface deformation dynamics in permafrost regions.
To validate the reliability of the models, the research team conducted comprehensive comparative analyses, systematically comparing InSAR-derived deformation, model-simulated deformation, and tiltarm-measured surface deformation data. The results showed that the improved model exhibited high consistency with the measured data during the heave phase and the fully frozen winter period, strongly confirming the model’s accuracy. However, the team also observed some discrepancies between the model’s results and the measured data during the thawing and subsiding phase. These differences are likely due to uneven migration of unfrozen water within the active layer during the autumn freezing process. This finding not only provides valuable scientific evidence for a deeper understanding of surface deformation dynamics in permafrost regions but also highlights the urgent need for more in-situ surface deformation data, which will provide a solid theoretical foundation and validation conditions for future model development.
Comparison of InSAR-derived Deformation, Model-simulated Deformation, and Tiltarm-measured Surface Deformation
The research was carried out with Lanzhou University as the leading institution, with PhD student Fan Chengyan as the first author, and Professors Mu Cuicui and Liu Lin from The Chinese University of Hong Kong as co-corresponding authors. Mr. Zhang Tingjun provided important support for the research.
