The century the climate stops waiting.

The planet has already warmed about 1.3°C above the 1850–1900 average, measured over the past decade. The year 2024 reached roughly 1.55°C, the first calendar year above 1.5°C in the modern record. A single hot year is not the same as crossing the 1.5°C long-term threshold, but the pace matters: the rate of warming since 1970 is the fastest in at least 2,000 years, and each of the past ten years ranks among the ten warmest on record.

Atmospheric carbon dioxide now exceeds 425 parts per million, roughly 50 percent above preindustrial levels and higher than at any point in at least 800,000 years of ice-core evidence. Methane and nitrous oxide have risen in parallel. Fossil-fuel emissions remain near 40 billion tonnes of CO₂ per year, and the ocean, which absorbs more than 90 percent of the excess heat, hit record heat content again in 2024. This is the starting condition: not a future threat, but an ongoing disturbance with momentum already built into the system.

The Intergovernmental Panel on Climate Change (IPCC) organises 21st-century projections into five socioeconomic pathways. The low-emission paths hold warming near 1.4°C to 1.8°C by 2100. The middle path reaches about 2.7°C. The high paths reach 3.6°C to 4.4°C. These are averages, not ceilings; the high pathway's upper range exceeds 5°C.

The realistic baseline sits between the middle and the high middle. Climate Action Tracker, which scores actual policies rather than pledges, puts the world on track for roughly 2.7°C by 2100, with a range from 2.4°C to 2.9°C. Pledges and targets announced but not yet delivered would bend that toward 2.1°C. Full delivery of every net-zero pledge would hold it near 1.8°C. The gap between pledges and policy is the single most important number in the outlook: the world is currently on the policy path, not the pledge path.

Crossing the 1.5°C long-term average is likely within the early 2030s on current trajectories. Crossing 2°C becomes likely in the second half of the century if policies do not tighten substantially. The difference between 2°C and 3°C is not linear; it is the difference between a strained system and one in which several major subsystems begin to fail simultaneously.

Warming is already distributed unevenly. Land warms faster than ocean. The Arctic warms at roughly four times the global average, which is why Arctic sea ice has lost about 12 to 13 percent of its September extent per decade since 1979. Nights warm faster than days, winters faster than summers, and the upper troposphere faster than the surface. These asymmetries shape the secondary effects: they determine where crops fail first, where rivers shrink, and where humid heat crosses the limits of human physiology.

Global mean sea level has risen about 10 centimetres since 1993 and is accelerating at roughly 3.6 millimetres per year. Projections for 2100 run from about 0.3 metres under deep mitigation to 1.1 metres or more under high emissions, with low-confidence ice-sheet processes capable of pushing the upper bound toward 2 metres. Ocean acidification, driven by the same CO₂, has already increased surface-ocean acidity by about 30 percent since preindustrial times.

The climate system contains self-reinforcing loops that can amplify a given forcing. Several are already measurably active. The water-vapour feedback is the largest and fastest: warmer air holds more moisture, and water vapour is itself a greenhouse gas, so each degree of warming adds an amplifying response roughly equal to the original CO₂ forcing. This feedback is well understood and is already built into every projection.

The ice-albedo feedback runs close behind. Melting snow and sea ice exposes darker ocean and land, which absorb more sunlight and accelerate the melt. This is why Arctic amplification runs at four times the global rate. It weakens the jet stream and has been linked to more persistent summer weather patterns in the mid-latitudes.

Tipping points are thresholds beyond which a system reorganises itself and does not return on human timescales, even if forcing stops. Recent synthesis work by Armstrong McKay and colleagues places several of them inside the range the planet may reach this century.

Greenland's ice sheet has a central threshold near 1.5°C and a range from 0.8°C to 3.0°C. Its collapse would unfold over centuries to millennia but would commit the world to roughly 7 metres of sea-level rise. The West Antarctic ice sheet sits in a similar band, from 1°C to 3°C and centred near 1.5°C. Coral reefs sit at 1.0°C to 2.0°C. The Amazon's dieback threshold centres near 3.5°C but interacts with deforestation, which can lower the effective threshold. Abrupt permafrost thaw centres near 4°C.

The Atlantic Meridional Overturning Circulation (AMOC) has a wide uncertainty range centred near 4°C; a collapse would cool parts of northwest Europe, shift tropical rainfall belts, and raise sea levels along the US east coast. Several recent studies argue the AMOC may be weaker than models assumed, which would pull its risk forward.

The remaining carbon budget is the binding constraint. To keep a 50 percent chance of holding warming to 1.5°C, the world can emit roughly 200 to 300 billion tonnes of CO₂ from 2025 onward, about five to eight years at current rates. The budget for 2°C is roughly 900 to 1,000 billion tonnes, about 25 years at current rates. Every year without deep cuts spends both budgets at once.

The technological path is no longer the primary obstacle. Solar and wind are now the cheapest new electricity in most markets; batteries and electric vehicles follow steep learning curves; heat pumps and electrolysers are scaling. The obstacle is speed. Holding 1.5°C requires global emissions to fall about 43 percent by 2030 and reach net zero near 2050. Holding 2°C allows net zero near 2070 but still requires a halving by 2050. Current policies do neither.

Carbon dioxide removal matters at the margin, not the centre. Residual emissions from aviation, agriculture, and industry must be balanced by removals in any net-zero pathway, but removals cannot substitute for cutting the 40 billion tonnes currently emitted each year. Adaptation runs in parallel: coastal defence, heat-resilient cities, drought-tolerant crops, early-warning systems, and cooling access. The adaptation bill is smaller than the damages it prevents, and it compounds because every tenth of a degree avoided lowers the adaptation burden for every subsequent decade.