› Permafrost Thaw & Active Layer Depth Simulator Back
Permafrost / Climate

Permafrost Thaw & Active Layer Depth Simulator

Predict the summer active layer depth in permafrost terrain using the Stefan equation. Change the mean annual air temperature, thawing degree days, soil type, ice content, vegetation cover, snow depth and warming scenario, and see the present-day active layer, ground temperature, future projection and permafrost carbon feedback (CO2eq emissions) in real time.

Parameters
Mean annual air temp MAAT
°C
Permafrost indicator. Below -2 °C: continuous permafrost
Thawing degree days TDD
°C·day
Sum of positive daily temperatures. Subarctic: 500-2000
Soil type
Sets thermal conductivity k, heat capacity, density
Ice content theta_ice
%
Volumetric ice fraction of the soil
Surface vegetation cover
%
Moss / lichen / shrub cover (summer insulation)
Winter snow depth
cm
Insulating blanket that raises winter ground temperature
ΔT warming
°C
IPCC AR6 SSP2-4.5 ~ +2 °C; SSP5-8.5 ~ +4 °C
Forecast horizon
year
How many years ahead to project the active layer
Results
—
Active layer (cm)
—
Mean ground T (°C)
—
Active layer in 50 yr (cm)
—
Δ active layer (cm)
—
Carbon release (kg C/m²)
—
CO₂eq emission (kg/m²)
—
Permafrost cross-section — seasonal animation

In summer the active layer thaws downward from the surface; in winter it refreezes. The green band is vegetation, the white band is snow, and CH4 bubbles rise from the thawed zone.

Active layer depth vs MAAT
Soil thermal property comparison
Theory & Key Formulas

$$h_{active} = \sqrt{\frac{2\,k_{soil}\cdot \text{TDD}}{L_v}},\quad L_v = \theta_{ice}\,\rho_{ice}\,L_f$$

k_soil: soil thermal conductivity [W/m/K]; TDD: thawing degree days in K·s; L_v: volumetric latent heat of fusion [J/m³] which depends on ice content. The classical analytical solution of the Stefan freezing/thawing problem.

$$h' = h_{active}\cdot(1-0.3\,V_{cov}),\quad T_g = T_{air}+5\cdot 0.4\cdot \frac{d_{snow}}{300}$$

V_cov: vegetation cover (up to 30% summer shading); d_snow: snow depth in cm (up to 40% winter insulation raising mean ground temperature). Surface-process corrections.

$$\Delta C = C_{perm}\cdot \frac{\Delta h}{1\,\text{m}}\cdot \eta_{decomp},\quad \text{CO}_2\text{eq} = \Delta C\cdot \text{GWP}_{\text{CH}_4}$$

C_perm ≈ 50 kg C/m² (typical permafrost soil carbon), η_decomp ≈ 10%, GWP_CH4 ≈ 28 (100-yr). First-order estimate of the permafrost carbon feedback.

Permafrost thaw and active layer depth — climate change response

🙋
So permafrost is just ground that stays frozen all year up north, right? But what is the "active layer"? What is "active" about it?
🎓
Good question. Permafrost is defined as ground that stays at or below 0 °C for at least two consecutive years. It covers about 24% of the Northern Hemisphere — Siberia, Alaska, northern Canada, and the Tibetan Plateau. But the surface is not frozen solid year-round. In summer, the top few tens of centimeters to several meters thaws, and in winter it refreezes. That summer-thaw / winter-freeze layer is the "active layer". Plants grow there, microbes decompose organic matter there. Everything biological in permafrost terrain happens inside this thin lid.
🙋
OK. So how is the active layer depth set? The tool uses "thawing degree days" instead of plain mean annual temperature.
🎓
That is the heart of geocryology. The key is the Stefan equation, h = sqrt(2 k TDD / L_v), which is surprisingly simple. TDD (Thawing Degree Days) is the integral of "daily temperature above 0 °C" over the summer — it captures the cumulative summer warmth. k is the soil thermal conductivity, and L_v is the volumetric latent heat needed to melt the pore ice. So the question is: given a heat budget, how deep can melting penetrate before the latent heat soaks it all up? That balance — heat conduction vs. latent heat — is what really matters, not the mean annual temperature on its own.
🙋
If I switch the soil type to peat, the active layer becomes very shallow. And bedrock makes it much deeper. That is just the conductivity?
🎓
Exactly. Peat is organic and porous, with k as low as 0.5 W/m/K. Dense bedrock has k = 3.0, six times higher. For the same summer heat, bedrock conducts heat down much faster and thaws deeper. That is why peatlands in the Siberian baidaras have active layers of only 30-50 cm, while gravelly slopes in interior Alaska can thaw to 1.5-2 m. Vegetation and snow also matter a lot. Push vegetation cover to 100% and the active layer drops by up to 30%; push snow to 300 cm and the winter cold is blocked and mean ground temperature climbs. Moss in summer and a thick snow blanket in winter are the real guardians of permafrost.
🙋
What is the "CO2eq emission" stat in the bottom right? Does deepening the active layer release CO2?
🎓
This is the famous "permafrost carbon feedback", one of the most-watched tipping points in climate science. Permafrost stores roughly 1,600 Gt of organic carbon — about half of all soil carbon on Earth — locked in ice. Mammoth carcasses included. When warming deepens the active layer, that previously frozen organic matter is exposed to microbes for the first time in thousands of years, and is released as CO2 and CH4. CH4 has 28 times the warming potential of CO2 over 100 years, so more emissions accelerate warming, which deepens the active layer further. A textbook positive feedback. Push ΔT to 4 °C in the tool and you'll see the 50-year active layer jump and the CO2eq emission shoot up.
🙋
So this is not just about climate — roads and buildings in the Arctic must be affected too. What kind of damage has actually happened?
🎓
It is serious. When the active layer deepens or the permafrost itself thaws, the ice in the soil melts to water, the soil loses volume, and the ground subsides. This process is called "thermokarst". The Trans-Alaska Pipeline is held off the ground on thousands of vertical piles with heat pipes precisely to keep the active-layer / permafrost interface stable, and even then pile settlement is a constant maintenance issue. In 2020 a fuel tank in Norilsk, Russia split open when permafrost subsidence undermined its foundation, spilling 21,000 tonnes of diesel into rivers. In Yakutsk, Fairbanks and many Inuit villages, houses are tilting, roads are buckling and airport runways are cracking. IPCC AR6 explicitly lists permafrost thaw as a climate-system tipping element.

Frequently asked questions

This tool uses the Stefan equation h = sqrt(2 k TDD / L_v), where h is the active layer depth [m], k is the soil thermal conductivity [W/m/K], TDD is the thawing degree days converted to seconds [K·s], and L_v is the volumetric latent heat of fusion [J/m³]. L_v is computed from the volumetric ice content θ_ice, the ice density and the latent heat of fusion as L_v = θ_ice·ρ_ice·L_f. The tool then applies vegetation and snow insulation corrections, and uses a warming scenario to raise TDD and mean air temperature for the future projection.
Permafrost stores about 1,600 Gt of organic carbon worldwide — roughly half of all soil carbon — kept frozen for thousands of years. As warming deepens the active layer, organic matter that was previously frozen is exposed to microbial decomposition and released as CO₂ and CH₄. CH₄ has about 28 times the warming potential of CO₂ over a 100-year horizon, so additional emissions accelerate warming and deepen the active layer further. This forms a positive feedback loop. This tool approximates that loop to first order by estimating the carbon released from the additional thaw and converting it to CO₂-equivalent.
Summer active layer depth is set by how much surface heat reaches the soil. Mosses, lichens and low shrubs shade the ground, lower the surface temperature and reduce heat flux, keeping the active layer shallower. This tool assumes a maximum reduction of about 30% at full vegetation cover. Winter snow does the opposite: snow has very low thermal conductivity and acts as a blanket that traps heat in the ground, weakening winter refreezing and raising the mean annual ground temperature. The tool models snow as providing up to about 40% insulation at 300 cm.
The Stefan equation is a classical 1D quasi-steady analytical solution that captures the order of magnitude of observed Siberian and Alaskan active layers (typically 30 to 200 cm). It does not include slope-dependent insolation, vertical soil moisture profiles, year-to-year weather variability, groundwater flow, or soil heterogeneity. For infrastructure design, engineers typically start with simple Stefan-style estimates like this tool, then validate with field thermistor strings, thermokarst surveys, and transient 1D-3D models such as CryoGrid or GIPL2.

Real-world applications

Arctic infrastructure design: The Trans-Alaska Pipeline (800 miles) is held above the ground on thousands of vertical heat-pipe-cooled piles, designed so the active-layer / permafrost boundary does not migrate. Tools like this one are used in early-stage design for pile embedment depth, foundation frost-jacking countermeasures, and the thickness of road insulation layers. The same considerations apply to railways (BAM line), airport runways and gas-field developments such as Yamal.

Climate models and carbon-cycle research: Earth System Models in CMIP6 / IPCC AR6 do not yet fully represent the permafrost carbon feedback. Specialised 1D-3D codes such as CryoGrid, GIPL2 and CLM-Permafrost compute active layer dynamics and carbon decomposition, combined with satellite observations (GRACE for groundwater, Sentinel-1 SAR for surface displacement) to estimate future CO₂ / CH₄ release. This tool is an entry point: a way to build intuition for how sensitive the Stefan-equation prediction is to its inputs.

Thermokarst and subsidence risk: Since 2014 giant craters (30 m wide) — most likely methane-driven blowouts — have been appearing on Russia's Yamal Peninsula. The 2020 Norilsk diesel spill (21,000 tonnes) was caused by permafrost subsidence under a fuel tank. For villages, pipelines and oil-tank farms, projecting the future active layer with tools like this is the first step in selecting heat pipes, insulation layers and reinforced foundations.

Indigenous communities and food security: In Nunavut and across Siberia, indigenous communities have stored reindeer meat and fish for generations in "ice cellars" cut into permafrost. As the active layer expands, cellar temperatures rise and traditional food storage fails. Combined with changes in caribou wintering grounds, rapid drainage of thermokarst lakes, and coastal erosion (5-20 m per year on the Arctic coast), permafrost thaw is reshaping daily life. Even simple Stefan-style estimates can contribute to long-term regional outlooks.

Common misconceptions and pitfalls

The most common misconception is that "permafrost = frozen to the surface year-round". In reality the upper few tens of centimeters to a few meters thaw every summer (that is the active layer), and only the layer below stays permanently frozen. In the continuous permafrost zone along the Arctic coast the permafrost layer itself can be hundreds of meters to 1,500 m thick, but in the discontinuous zone to the south it is patchy and only 10-50 m thick — and that southern zone is the first to disappear under warming. This tool only computes the active layer; predicting when the permafrost layer itself will disappear requires transient heat-conduction modelling over hundreds to thousands of years.

The second pitfall is treating the Stefan equation as a precise predictor. It is a 1D, quasi-steady, latent-heat-dominated idealisation. Real permafrost involves groundwater flow, vertically heterogeneous ice content, year-to-year weather noise, slope-dependent insolation, and vegetation feedbacks (wetland formation, etc.). At many Siberian observation sites the difference between observed and Stefan-predicted active layer reaches ±30%. Treat the values from this tool as order-of-magnitude estimates and sensitivity guides — for real projects, combine on-site boreholes, thermistor strings, satellite InSAR and a CryoGrid-class model.

The third pitfall is the assumption that "thawing permafrost is a local problem". In fact, the permafrost carbon feedback is global. Schuur et al. (Nature 2015) estimated that 4 °C warming could release 130-160 Gt CO₂eq by 2100 — several years' worth of current fossil-fuel emissions (~40 Gt/yr). CH₄ is short-lived but high-GWP, and abrupt thermokarst-lake and crater emissions can be locally measurable. Permafrost also stores ancient viruses and bacteria (Pithovirus was reactivated from Siberian permafrost in 2014) and about half of all Arctic mercury, so thaw has biosecurity and toxicology consequences beyond the carbon cycle.

How to Use

  1. Input mean annual air temperature (°C) and thawing degree-days (sum of daily temperatures above 0°C during summer) for your location
  2. Specify volumetric ice content (%) and vegetation cover type (bare soil, moss, tundra grasses) which affect surface albedo and thermal conductivity
  3. Run the Stefan equation solver to calculate active layer depth, predict 50-year thaw progression, and estimate carbon release from thawing organic soils

Worked Example

For a tundra site in northern Alaska: mean annual air temperature = -8°C, thawing degree-days = 800 K-days, volumetric ice content = 35%, vegetation cover = moss. The simulator calculates active layer depth = 62 cm using the Stefan equation (depth proportional to √TDD where TDD = thawing degree-days). With 1°C warming over 50 years, active layer deepens to 78 cm (delta = +16 cm). Organic soil carbon release = 3.2 kg C/m² (approximately 11.7 kg CO₂eq/m² accounting for CO₂ and CH₄ emissions).

Practical Notes

  1. Sparse vegetation (lichen, dwarf shrubs) insulates permafrost better than bare bedrock; dense moss mats reduce active layer thaw by 5–8 cm compared to unvegetated surfaces
  2. High ice content (40%+) in fine silts amplifies subsidence risk; ground settlement of 15–20 cm per meter of thaw depth damages infrastructure like pipelines and building foundations
  3. Validation data: Stefan predictions agree within ±10% of field measurements from CALM network (Circumpolar Active Layer Monitoring) sites in Siberia and Canadian Arctic