Pacific Cold Tongue Bias in CMIP5/CMIP6 Linked to Shifts in Extratropical Subduction Zones

Abstract The equatorial Pacific cold tongue (CT) bias is a systematic sea surface temperature (SST) bias, persisting throughout all generations of comprehensive climate models. Recent works have suggested that extratropical SST biases contribute to the CT bias, mediated by the wind-driven ocean overturning circulation in the Pacific. However, as shown here, the Northern and Southern Hemispheric “eastern exchange windows,” which are the dominant extratropical sources for water upwelling in the equatorial Pacific, are not characterized by cold SST biases. Here, we explore these links using Lagrangian back-trajectory analysis in 32 models participating in phases 5 and 6 of the Coupled Model Intercomparison Project (CMIP5/CMIP6), and four ocean reanalyses. Consistent with previous works, equatorial Pacific upwelling is found to be sourced primarily from late-winter subduction in the extratropical eastern exchange windows. We also find that climatological dynamical fields dominate the probability distribution maps linking extratropical subduction with upwelling in the CT. Variations across CMIP5/CMIP6 models and between CMIP5/CMIP6 models and reanalyses consistently point to poleward shifts of the eastern exchange windows into colder extratropical waters as a likely key contributing factor to the CT bias, which, due to the strong meridional SST gradients in the extratropics, can drive CT biases regardless of extratropical SST biases. Consistent with previous works, cold biases in the northern extratropics, which partially overlap with the northeastern exchange window, also contribute to the CT bias. Trajectory duration varies considerably across models and between models and reanalyses, but it is not consistently related to the CT bias. Significance Statement Systematic surface temperature biases in the equatorial Pacific, which are linked to systematic tropical precipitation biases, limit the ability of climate models to make reliable predictions. Recent works suggest that these systematic biases are related to temperature biases in water that sinks in the extratropics, travels equatorward, and resurfaces in the equatorial Pacific. Here, we test this link in modern climate models and in observational datasets by directly calculating the trajectories of water parcels along their journey from the extratropics to the equator. We find that in modern climate models, the extratropical sinking regions are shifted poleward toward colder water, ultimately leading to cooler rising water at the equatorial Pacific. The results, therefore, point to ocean circulation as a critical culprit in the systematic equatorial Pacific temperature biases.