The planet's permafrost — 20 million square kilometres of permanently frozen ground across Siberia, Canada, Alaska, and northern Europe — is not a passive victim of climate change. It is an active participant. Gustaf Hugelius, professor at Stockholm University's Bolin Centre for Climate Research, lays out the mechanism with uncomfortable clarity: 1,500 gigatons of carbon sit locked in frozen soil, three times the carbon stored in every living plant on Earth. As warming thaws the ground, microbes consume ancient organic matter and exhale CO2 and methane. That warms the atmosphere further, which thaws more permafrost. The loop feeds itself. The irreversibility is the critical detail. When permafrost thaws, underground ice melts, water drains, and the surface collapses into thermokarsts — lakes, wetlands, gullies with fundamentally different hydrology and thermal properties. This landscape does not refreeze on any human timescale. Several million square kilometres along the southern fringes of the permafrost zone are already doomed, their threshold exceeded a decade or more ago, even if global temperatures stabilised today. Northern Sweden and northern Finland are among the areas already past the point of return. The emissions math is staggering. Between 200 and 300 gigatons of CO2 equivalents could escape this century under warming scenarios of 1.8°C to 3.6°C — many times the EU's current annual emissions. In higher warming scenarios, permafrost emissions could rival those of the United States or China. Whether the released gas is CO2 or methane depends on whether thawed landscapes go wet or dry: dry conditions produce CO2 (long-lived, centuries-scale threat), wet conditions produce methane (30 times more potent short-term, breaks down in 10-15 years). Researchers still cannot predict the wet-dry split with confidence. The system will not tip all at once — the region is too vast for a single global threshold short of 6°C warming. Instead, local tipping points are firing in succession, each one permanently converting frozen ground into an emitting landscape. Heatwaves and wildfires in Siberia and Canada are accelerating the process. Hugelius reports that since his first Arctic expedition in 2007, warm summers now reliably produce new thermokarst lakes within years. The economic ledger is almost entirely negative. Purported benefits — wider agricultural frontiers, exposed mineral resources — are illusory. Thawing ground becomes unstable, making resource extraction more expensive and dangerous. Half of Russia's oil and gas production sits on permafrost that will degrade within decades. Agriculture cannot simply move north onto newly thawed, forested soils. Russia, Canada, and Alaska face infrastructure damage with no offsetting gains. No geoengineering fix exists at the required scale. The only lever is emissions reduction: every degree of warming prevented saves 4 million square kilometres of permafrost. This is the brutal arithmetic of a feedback loop — delay makes the problem self-amplifying. The more the Earth system is disturbed, the more it emits autonomously, compounding the human challenge. Hugelius frames the stakes across generations: the damage being locked in now will force 5 to 10 generations to manage autonomous emissions from a system humans can no longer control. The science on the process and the feedback is virtually certain. The uncertainty is in the rate and the methane-vs-CO2 split — details that determine severity, not whether the problem exists.
