We review the evidence for a putative early 21st-century divergence between global mean surface temperature (GMST) and Coupled Model Intercomparison Project Phase 5 (CMIP5) projections. We provide a systematic comparison between temperatures and projections using historical versions of GMST products and historical versions of model projections that existed at the times when claims about a divergence were made. The comparisons are conducted with a variety of statistical techniques that correct for problems in previous work, including using continuous trends and a Monte Carlo approach to simulate internal variability. The results show that there is no robust statistical evidence for a divergence between models and observations. The impression of a divergence early in the 21st century was caused by various biases in model interpretation and in the observations, and was unsupported by robust statistics.
Lewandowsky, S., Cowtan, K., Risbey, J. S., Mann, M. E., Steinman, B. A., Oreskes, N., & Rahmstorf, S. (2018). The ‘pause’ in global warming in historical context: (II). Comparing models to observations. Environmental Research Letters, 13(12), . https://doi.org/10.1088/1748-9326/aaf372