Observational Attribution-Constrained Projections of Area-Averaged Annual Mean Temperature Change over China

Accurate regional temperature projections in China are vital for climate change mitigation and adaptation. Since some of the Coupled Model Intercomparison Project Phase 6 (CMIP6) climate models are identified as "too warm" with a tendency to overestimate future warming, this study introduces an observational attribution-constrained projection method to reduce uncertainties in regional temperature projections. By applying this method to the CMIP5 and CMIP6 models, we establish a robust linkage between observed and projected temperature changes across China. Using the optimal fingerprinting technique, we demonstrate the dominant influence of anthropogenic activities on regional temperature changes. Through application of the optimal estimates of scaling factors derived from this method, we constrain future temperature projections. Under high-emissions Representative Concentration Pathway (RCP8.5) and Shared Socioeconomic Pathway scenarios (SSP5-8.5), the unconstrained annual mean temperature changes for 2081–2100 relative to 1995–2014 are projected as 4.47°C (5th–95th percentile range: 3.38–6.23°C, hereafter the same) by CMIP5 and 5.24°C (3.60–7.76°C) by CMIP6. After constraint implementation, these projections decrease to 3.90°C (2.98–4.48°C) for CMIP5 and 4.35°C (3.49–5.11°C) for CMIP6, with reduced uncertainty ranges compared to original projections. The constrained method improves consistency between CMIP5 and CMIP6 simulations, reducing inter-model uncertainties and enhancing the reliability of China’s temperature change projections.