Abstract
Long-term measurements of permafrost temperatures do not provide a complete picture of the Arctic subsurface thermal regime. Regions with warmer permafrost often show little to no long-term change in ground temperature due to the uptake and release of latent heat during freezing and thawing. Thus, regions where the least warming is observed may also be the most vulnerable to permafrost degradation. Since direct measurements of ice and liquid water contents in the permafrost layer are not widely available, thermal modeling of the subsurface plays a crucial role in understanding how permafrost responds to changes in the local energy balance. In this work, we first analyze trends in observed air and permafrost temperatures at four sites within the continuous permafrost zone, where we find substantial variation in the apparent relationship between long-term changes in permafrost temperatures (0.02–0.16 K yr−1) and air temperature (0.09–0.11 K yr−1). We then apply recently developed Bayesian inversion methods to link observed changes in borehole temperatures to unobserved changes in latent heat and active layer thickness using a transient model of heat conduction with phase change. Our results suggest that the degree to which recent warming trends correlate with permafrost thaw depends strongly on both soil freezing characteristics and historical climatology. At the warmest site, a 9 m borehole near Ny-Ålesund, Svalbard, modeled active layer thickness increases by an average of 13 ± 1 cm K−1 rise in mean annual ground temperature. In stark contrast, modeled rates of thaw at one of the colder sites, a borehole on Samoylov Island in the Lena River delta, appear far less sensitive to temperature change, with a negligible effect of 1 ± 1 cm K−1. Although our study is limited to just four sites, the results urge caution in the interpretation and comparison of warming trends in Arctic boreholes, indicating significant uncertainty in their implications for the current and future thermal state of permafrost.
Introduction: The Latent Heat Paradox
Monitoring the Arctic subsurface is notoriously difficult. While boreholes provide ground temperature readings, these data points often fail to capture the full "thermal state" of permafrost. The primary complication is latent heat: as permafrost approaches the freezing point, incoming energy is consumed by the phase change of melting ice rather than increasing the measurable temperature.
This creates a paradox where regions showing the least temperature change may be the most vulnerable to rapid structural degradation. Because direct measurements of subsurface ice and water content are scarce, there is a significant gap in understanding "unseen" thawing. This study addresses this gap by using Bayesian inverse modeling to infer hidden variables—such as latent heat and active layer thickness (ALT)—from existing temperature records.
Methods: Bayesian Inversion and Physical Modeling
The researchers integrated a physical heat transfer model with a robust statistical framework:
- Physical Model: A 1D transient heat conduction model accounting for phase changes via a Soil Freezing Characteristic Curve (SFCC), which defines the ratio of unfrozen water to ice at sub-zero temperatures.
- Bayesian Inversion: To manage uncertainties in soil properties and surface conditions, the study utilized Ensemble Kalman Sampling (EKS). EKS is a parallelizable, derivative-free method that efficiently approximates the "posterior" distribution of the system's state, making it more computationally feasible than traditional Markov Chain Monte Carlo (MCMC) methods for complex simulations.
- n-Factors: Time-dependent "n-factors" were used as tunable parameters to bridge the gap between air temperature and actual ground surface temperature, accounting for the insulating effects of snow and vegetation.
Study Sites and Data
The analysis focused on four distinct sites within the continuous permafrost zone, representing a range of climatic and geological conditions.
The sites include:
- Samoylov Island (Siberia): Cold, ice-rich polygonal tundra.
- Bayelva (Svalbard): Relatively warm, silty/clayey coastal site.
- Parson’s Lake (Canada): Shrub tundra characterized by high soil heterogeneity.
- Utqiaġvik/Barrow (Alaska): Cold, coastal site with variable ice content.
The study utilized over 10 years of borehole temperature records and bias-corrected ERA5-Land reanalysis data for atmospheric forcing.
Key Results and Insights
The study revealed that permafrost temperature trends alone are a deceptive metric for Arctic health.
1. The Masking Effect of Phase Change
In warmer permafrost zones like Svalbard, ground temperatures often remain "stuck" near 0°C. This "zero curtain" effect occurs because energy is diverted into melting ice. Consequently, these sites may appear stable in temperature records while undergoing significant internal melting and active layer deepening.
2. Divergent Thaw Sensitivity
The sensitivity of the Active Layer Thickness (ALT) to warming varies by an order of magnitude depending on the site's historical climatology and soil texture:
- Bayelva (Warm Site): The modeled active layer deepened by 13 ± 1 cm for every 1 K rise in mean annual ground temperature.
- Samoylov (Cold Site): Sensitivity was negligible (1 ± 1 cm/K), as energy primarily contributed to the sensible heating of very cold ground rather than phase change.
3. Soil Composition as a Driver
Soils with high silt or clay content (e.g., Barrow, Bayelva) possess a wider "mushy zone" where ice and water coexist. This increases latent heat capacity and enhances the thermal inertia of the permafrost compared to sandier sites like Samoylov.
Limitations and Future Directions
While providing a powerful new lens for permafrost analysis, the model has limitations:
- Static Hydrology: The assumption of constant soil water content ignores the dynamic movement of water during thaw, which affects thermal conductivity.
- Surface Physics Proxies: The use of "n-factors" simplifies complex processes like evapotranspiration and snow insulation.
- Gaussian Assumptions: The EKS method assumes Gaussian distributions, which might underestimate uncertainty in highly non-linear systems.
Future work aims to scale this Bayesian approach to circum-Arctic maps, potentially integrating satellite-derived ground temperatures to better constrain global permafrost models.
Summary updated