Climate change projections
Pro Research Analysisby 
Searched over 200M research papers
Climate Change Projections: Global Warming Estimates and Model Uncertainties
Projected Global Warming Ranges for the 21st Century
Climate change projections for the 21st century show a range of possible global surface warming outcomes, depending on the models and emission scenarios used. Some projections estimate warming between 1.3°C and 8.0°C by 2100, but recent research suggests that the most extreme scenarios are unlikely. More realistic projections, especially those based on moderate emission pathways like SSP2-4.5, indicate that global warming is likely to remain below 2.5–3.0°C, and possibly even under the 2.0°C threshold set by the Paris Agreement, if moderate adaptation policies are followed .
Model Ensembles and Sources of Uncertainty
Climate projections are typically generated using ensembles of global and regional climate models. However, these ensembles often have uneven representation, with some models and scenarios being favored over others. The main source of uncertainty in these projections is the choice of the driving global climate model, rather than the emission scenario or model resolution, especially for near-future changes in temperature and precipitation . Additionally, similarities among models can bias projections and underestimate uncertainty. When accounting for model similarity, the projected global mean temperature rise by the end of the century is slightly lower than previously estimated, especially under high emission scenarios .
The Role of Observational Constraints
Recent advances have focused on constraining climate projections using historical temperature records and observed warming trends. These observational constraints help narrow the uncertainty in future warming estimates by about 50%, making projections more robust and excluding the lower end of previous warming estimates Ribes2021Shiogama2016. When models are weighted based on their ability to reproduce observed trends, the upper bounds of projected warming are also reduced, leading to more reliable forecasts .
Probabilistic and Machine Learning Approaches
New methods, such as neural network-based climate model substitutes, have improved the efficiency and accuracy of large climate model ensembles. These approaches use observed surface warming and ocean heat uptake to constrain key parameters, resulting in probability distributions for future warming that are consistent with observations. However, some projections still show warming exceeding the ranges previously reported by the IPCC, especially if higher climate sensitivity values are assumed .
Implications for Policy and Adaptation
Observationally informed projections suggest that achieving global temperature stabilization targets may require more aggressive greenhouse gas emissions reductions than previously calculated. In some cases, the mean projected warming is about 0.5°C higher than raw model projections, with a narrower range of uncertainty . This highlights the importance of using updated, observation-constrained models for policy planning and adaptation strategies.
Regional and Upper Atmosphere Projections
While most projections focus on surface temperature, some studies extend to regional impacts and changes in the upper atmosphere. For example, under moderate emission scenarios, significant cooling and density decline are expected in the thermosphere, with notable spatial variations due to changes in Earth's magnetic field . Regional projections, such as those for Spain and the Balearic Islands, show that uncertainty is dominated by the choice of global model rather than regional model or scenario .
Conclusion
Current research indicates that realistic climate change projections for the 21st century are likely to result in global warming below 2.5–3.0°C, with a strong possibility of staying under 2.0°C if moderate adaptation measures are implemented. Advances in model weighting, observational constraints, and machine learning have significantly reduced uncertainty in these projections. However, achieving climate stabilization targets may require more ambitious emissions reductions than previously thought, emphasizing the need for continued refinement of climate models and adaptation strategies Scafetta2023Ribes2021Shiogama2016+1 MORE.
Sources and full results
Most relevant research papers on this topic