A comprehensive analysis comparing observed climate change since 1990 against the projections of climate models available at that time finds that the models accurately predicted the magnitude and geographic distribution of temperature increases, sea level rise, arctic ice loss, and extreme precipitation intensification. The analysis directly addresses claims that climate models are unreliable, finding instead that model predictions have been consistently accurate at the global scale and increasingly accurate at regional scales as models have improved.
The validation has important implications for the credibility of future projections. If models accurately predicted observed changes over 35 years, there is stronger scientific basis for trusting their projections about future changes under different emissions pathways. The same analysis finds that economic and energy system models used in policy analysis have been significantly less accurate than climate models, suggesting that the uncertainty most affecting climate policy decisions relates to human behavior rather than physical climate science.
