Unlike the current distribution map, this data constructs a machine learning comprehensive evaluation model that integrates "measured disaster disturbance information" and "potential environmental risk factors", achieving quantitative risk zoning for uninspected road sections and potential high-risk areas. The data is characterized by an instability index (0.0-1.0) to indicate the degree of risk, revealing the spatial coupling relationship between the thermal stability of permafrost, topographical and hydrological conditions, and human engineering disturbances (such as road heat island effect and pipeline active heat source). This achievement provides key data support for risk identification, preventive maintenance, and route planning of linear engineering in cold regions.
The core field of the dataset is the instability index (0-1.0), which is divided into five levels according to the degree of risk:
0-0.20 (extremely stable/low-risk): mainly distributed in the seasonally frozen soil area south of the southern boundary (SLLP).
0.20-0.40 (high stability/low risk): Belongs to weakly sensitive areas, mostly found in mountain ridges with low ground temperature (<-1.5 ℃) and good drainage.
0.40-0.60 (moderate stability/moderate risk): Key warning area, mainly distributed in the degradation transition zone between the southern boundary of permafrost in the 1970s and 1990s.
0.60-0.80 (low stability/high risk): The project is located in an unstable zone, adjacent to the main roads of national and provincial highways, and prone to thermal and melting disasters.
0.80-1.00 (extremely low stability/extremely high risk): Key intervention areas, highly focused on transportation hubs along roads (railways, highways), the China Russia crude oil pipeline (CRCOP), and large areas of discontinuous permafrost in the north.
Independently generated, integrating multiple field scientific expedition records and machine learning models to obtain prediction results.
Comprehensively utilizing machine learning algorithms to invert the weights of ground temperature (LST), ground temperature (MAGT), vegetation cover (NDVI), terrain (slope direction), and engineering thermal disturbance factors. By integrating measured freezing damage points for supervised learning, a continuous instability index model is constructed to achieve the transformation from point like status quo to surface level risk potential.
Input parameters: Integrated data on the transition of the southern boundary between the 1970s and 1990s in the permafrost region of Northeast China.
Disturbance simulation: Modeling the physical mechanisms of the heat island effect caused by passive heat collection on highways and the local melting cylinder structure caused by active heat dissipation on crude oil pipelines.
Accuracy assessment: By conducting on-site verification of typical vulnerable areas such as Mohe Hub and Genhe Yakeshi degradation front, the accuracy of risk level classification is ensured.
(1) Accuracy and Scale: The data is presented at a scale of 1:100000 with a spatial resolution of 30 meters, ensuring precise representation within linear engineering buffer zones (highways, railways, pipelines).
(2) Model method: A comprehensive evaluation model integrating "measured freeze-thaw disaster disturbance information" and "potential environmental risk factors" is adopted. By calculating the instability index, the transition from qualitative description to quantitative evaluation can be achieved.
(3) Standard specification: The evaluation level is strictly divided into five gradients (0-0.2 to 0.8-1.0), with clear definitions for each level. For example, the high-risk area (0.8-1.0) was validated through the mechanism of hedging effects, while the high stability area (0.2-0.4) referred to physical indicators such as multi-year average ground temperature and soil drainage conditions.
(4) Verification and reliability: The risk distribution results show high consistency with the large discontinuous permafrost areas in the northern part of the Greater Khingan Range and the island shaped permafrost areas in the central and southern parts. We conducted on-site disturbance simulation verification specifically for typical corridors such as G111, G331, and the China Russia crude oil pipeline, ensuring the guiding value of the predicted results for preventive maintenance.
| # | number | name | type |
| 1 | 2022FY100700 | Survey of Permafrost Conditions and Freeze-Thaw Damage in the High-Latitude Regions of Northeast China | Basic Resource Survey Project |
This work is licensed under
CC BY 4.0 (Creative Commons Attribution 4.0 International License).
CSTR:11738.11.NCDC.NIEER.DB7264.2026
10.12072/ncdc.nieer.db7264.2026
