Evaluation of sea surface temperature interhemispheric dipole in CMIP6 historical simulations

The sea surface temperature (SST) interhemispheric dipole (SSTID) mode is an important multidecadal variability of global SST. It reflects hemispherically antisymmetric changes in SST between northern and southern hemispheres and contributes to north-south asymmetry of tropical land precipitation (TLP) on multidecadal timescales. This study evaluates performances of the 40 models from Coupled Model Intercomparison Project Phase 6 (CMIP6) historical experiments in capturing basic features of the SSTID and its physical connection to TLP. A systematic evaluation indicates that most of models successfully capture spatiotemporal features of the SSTID, while discrepancies also exist. Although the multi-model ensemble (MME) result successfully reproduces basic characteristics of the SSTID, it tends to overestimate spatial simulations, and there is room for improvement in its temporal simulations. To further explore the role of SSTID simulation in accurately reproducing the asymmetry of TLP on multidecadal timescales, we employ the conditional multi-model ensemble (CMME) method. Compared with the MME, the CMME effectively reproduces the TLP asymmetry on multidecadal timescales by correcting overly strong signals in the MME. This finding suggests that more reliable multidecadal north-south asymmetry of TLP simulations could be attributed to better performances in simulating the SSTID, which provides new insights for projection of TLP on multidecadal timescales.