Principle-based adept predictions of global warming from climate mean states.

Distinguishing anthropogenic warming from natural variability and reducing uncertainty in global-warming projections continue to present challenges. Here, we introduce a novel principle-based framework for predicting global warming from climate mean states that is based solely on carbon-dioxide-increasing scenarios without running climate models and relying on statistical trend analysis. By applying this framework to the climate mean state of 1980-2000, we accurately capture the subsequent global warming (0.403 K predicted versus 0.414 K observed) and polar warming amplification patterns. Our predictions from climate mean states of individual models not only exhibit a high map-correlation skill that is comparable to that of individual Coupled Model Intercomparison Project Phase 6 models for the observed warming, but also capture the temporal pace of their warming under the 1% annual CO2-increasing scenario. This work provides the first principle-based confirmation that anthropogenic greenhouse gases are the primary cause of the observed global warming from 1980-2000 to 2000-2020, independently of climate models and statistical analysis.