Every summer, heat diffuses down from a warmed ground surface into frozen soil, pushing the 0°C boundary — the base of the "active layer" — deeper before winter refreezes it. This simulator solves that heat-diffusion front on a depth grid: surface warming and insulation (snow/moss trap or block heat) set the boundary condition, and ground ice content sets the effective thermal conductivity a real permafrost monitoring network measures with borehole thermometers at fixed depths.
∂T/∂t = k · ∂²T/∂z²
active-layer depth = deepest z where T(z) > 0°C
early-warning trigger: any borehole reads T > 0°C
- Summer surface warming — how far above 0°C the surface swings each season; stronger warming drives the thaw front deeper, faster.
- Snow / moss insulation — higher values slow heat exchange in both directions, damping the seasonal swing that reaches the frozen ground.
- Ground ice content — ice-rich soil absorbs more energy per degree of thaw (latent heat), so the front advances more slowly but a deep thaw releases more meltwater when it does.
- Borehole sensors — the three depth probes on the chart flip to "ALERT" the instant their local temperature crosses 0°C, mirroring how real early-warning networks flag thaw before it is visible at the surface.
Real-world relevance: as waterlogged, newly-thawed soil goes anaerobic, microbes decompose the organic matter it holds and release methane — the "methane risk" readout tracks how much of the active layer has thawed into that waterlogged state, the same feedback loop early-warning monitoring networks are built to catch before it accelerates.