Read the observation period before the headline
Biskaborn and colleagues reported a global permafrost-temperature increase of 0.29 ± 0.12°C for the 2007–2016 reference decade. Their 2019 paper uses a network of boreholes, with temperatures measured near the depth of zero annual amplitude, where the seasonal temperature cycle is small. The number is neither an air-temperature rise nor the percentage of permafrost area that disappeared.
The full dataset contains 154 boreholes; 123 had adequate records for estimating decadal change. Of those, 71 warmed, 12 cooled and 40 were unchanged within measurement accuracy. Dividing 71 by 123 gives approximately 57.7% of those boreholes. It does not mean 57.7% of the world’s permafrost area warmed: observation sites are not equal-area samples.
Why the mean and the site count answer different questions
The published regional and global trends use area-weighted means and treatment of clustered sites. A simple count of red and blue dots on a map would give dense monitoring clusters excessive influence. A warming average can coexist with local cooling, just as a city’s average rainfall can increase while one gauge records a drier year.
The authors report warming of 0.39 ± 0.15°C in continuous permafrost and 0.20 ± 0.10°C in discontinuous permafrost over the reference decade. Ground temperatures respond to snow insulation and local surface conditions as well as air temperature. This helps explain why a nearby weather station cannot substitute for a borehole record.
Warming is not yet a carbon-emission measurement
A ground-temperature trend indicates a changing thermal state. To infer settlement, one also needs ground-ice and displacement evidence. To infer carbon emissions, one needs carbon availability and gas-flux measurements or an explicitly specified model. Moving directly from a temperature-colored map to tonnes of carbon skips both the physical and biological measurement steps.
These observations describe 2007–2016. They should not be labeled “current conditions” or extended linearly to 2026 without new measurements. The reported ± values belong to the study’s uncertainty treatment, not to every individual sensor or a forecast of future thaw.
Biskaborn et al. (2019), Permafrost is warming at a global scale, Nature Communications 10, 264. Published borehole results are summarized, not reprocessed. The 57.7% figure is 71/123 × 100; it describes the usable borehole sample only. Source checked 30 October 2026; analysis updated on that date.
What the definition does and does not say
Permafrost is defined by ground temperature and duration, not by permanent surface ice or a particular landform. The active layer can thaw every summer while deeper ground still meets the definition. One warm air-temperature season does not by itself establish that permafrost at a site has disappeared.
Match each question to a measurement
| Record | What it can show | Main limitation |
|---|---|---|
| Borehole temperature | Thermal state at sensor depths and dates | Sparse sites; equipment and local ground conditions matter |
| Active-layer thickness | Seasonal maximum thaw at a plot or grid | Does not map the full permafrost body |
| InSAR or repeat elevation | Surface displacement over mapped intervals | Subsidence is not uniquely caused by thaw |
| Field and lake mapping | Thermokarst forms, drainage and erosion at observed sites | Coverage and classification vary |
| Gas flux or soil carbon | Exchange or stocks under stated methods | Cannot be extrapolated across the Arctic without a model |
Thermokarst is an interpretation with evidence
When ground ice melts, volume loss can cause irregular subsidence, altered drainage and ponding. Calling a depression thermokarst should be supported by ground-ice context, repeat topography or field evidence. Similar shapes can result from erosion, sediment compaction or other processes.
A local change is not an Arctic-wide rate
Permafrost temperature, ice content, vegetation, snow, geology, wildfire history and hydrology vary sharply. A dramatic site photograph cannot establish regional area loss, average carbon release or infrastructure risk.
Carbon feedback claims need a chain of measurements
Frozen soils contain organic carbon. Thaw can make some material available for microbial decomposition, but carbon dioxide and methane fluxes depend on moisture, oxygen, substrate, vegetation and time. The sign and magnitude of a measured flux cannot be inferred from subsidence alone.
Evidence task
For a reported “permafrost collapse,” record coordinates, observation dates, ground temperature or ice evidence, elevation-change method, spatial resolution and one competing explanation. Then state whether the evidence supports thermal change, surface change, or both.
Sources and limits
- National Snow and Ice Data Center: frozen ground and permafrost.
- Global Terrestrial Network for Permafrost: borehole and active-layer monitoring context.
- IPCC AR6 WGI Chapter 9: ocean, cryosphere and sea-level evidence assessment.
Sources accessed 10 October 2026. This page supplies no live hazard map, engineering design value or site forecast.