资源环境科技发展态势分析平台

Tropical Pacific climate varies at interannual, decadal and centennial time scales, and exerts a significant influence on global climate. Climate model projections exhibit a large spread in the magnitude and pattern of tropical Pacific warming in response to greenhouse-gas forcing. Here, we show that part of this spread can be explained by model biases in the simulation of interannual variability, namely the El Nio/Southern Oscillation (ENSO) phenomenon. We show that models that exhibit strong ENSO nonlinearities simulate a more accurate balance of ENSO feedbacks, and their projected tropical Pacific sea surface temperature warming pattern is closely linked to their projected ENSO response. Within this group, models with ENSO nonlinearity close to observed project stronger warming of the cold tongue, whereas models with stronger than observed ENSO nonlinearity project a more uniform warming of the tropical Pacific. These differences are also manifest in the projected changes of precipitation patterns, thereby highlighting that ENSO simulation biases may lead to potentially biased projections in long-term precipitation trends, with great significance for regional climate adaptation strategies.