Each forecast method generates a distribution of cumulative CO₂ totals (2000–2050) via Monte Carlo sampling of its 95% CI. The fraction of samples falling between each pair of adjacent SSP cumulative benchmarks determines the scenario probability. The ensemble sits in the SSP4-3.4 to SSP2-4.5 corridor (65%/35%), though the Growth Rate method — with its higher cumulative mean — leans mostly SSP2-4.5 (83%), while RW-Quad 25yr places 100% of samples in SSP4-3.4.
| Method | Cumul. Mean | SSP1-2.6 | SSP4-3.4 ★ | SSP2-4.5 | SSP4-6.0 |
|---|---|---|---|---|---|
| Growth Rate Extrap. | 1,786 | 0% | 16% | 83% | 1% |
| Gaussian Process | 1,684 | 2% | 76% | 22% | 0% |
| RW-Quad 25yr | 1,672 | 0% | 100% | 0% | 0% |
| RW-Quad 20yr | 1,726 | 0% | 66% | 34% | 0% |
| Ensemble Average | 1,717 | 0% | 65% | 35% | 0% |
Emissions|CO2|Energy and Industrial Processes) ·
Observed: Global Carbon Budget 2025 (Friedlingstein et al.) for 2000–2024; 2025 = GCB preliminary estimate (38.1 GtCO₂) ·
globalcarbonbudget.orgEnter your details and Dr Ben McNeil will be in touch.