Historical cold biases exaggerate future elevation-dependent warming over the Tibetan Plateau

The Tibetan Plateau, often referred to as the third pole and Asian water tower, has warmed twice of the global mean since 1979. This fundamental phenomenon, known as Tibetan Amplification (TA), is projected to persist into the future, with important consequences on the Tibetan Plateau and the broader climate system. Yet, future magnitude of TA remains highly uncertain, varying widely among model projections from 1.2 to 1.8 by the end of the 21st century. This study demonstrates that historical surface temperature over the Tibetan Plateau serves as an effective constraint on projected TA. By attributing TA through a theoretical index and integrating reanalysis datasets with CMIP6 models, we find that cold biases in historical surface temperature exaggerate the positive contributions of albedo and lapse-rate feedbacks to TA, leading to inflated TA projections. Constraining models with historical surface temperature lessen TA ~15%, with uncertainty reduced 16%. Moreover, TA overestimation exhibits a robust elevation dependence, increasing by ~0.17% per 100 m. These results indicate that raw model output may overestimate future climate risks associated with TA.