Influence of Model Bias on Simulating North Atlantic Sea Surface Temperature During the Mid-Pliocene

Climate models generally underestimate the pronounced warming in the sea surface temperature (SST) over the North Atlantic during the mid‐Pliocene that is suggested by proxy data. Here we investigate the influence of the North Atlantic cold SST bias, which is observed in many climate models, on the...

Full description

Bibliographic Details
Published in:Paleoceanography and Paleoclimatology
Main Authors: Song, Zhaoyang, Latif, Mojib, Park, Wonsun, Zhang, Yuming
Format: Article in Journal/Newspaper
Language:English
Published: AGU (American Geophysical Union) 2018
Subjects:
Online Access:https://oceanrep.geomar.de/id/eprint/44136/
https://oceanrep.geomar.de/id/eprint/44136/4/Song_et_al-2018-Paleoceanography_and_Paleoclimatology.pdf
https://oceanrep.geomar.de/id/eprint/44136/2/palo20648-t-sup-0001-2018pa003397-t-figure_si-s01.docx
https://doi.org/10.1029/2018PA003397
Description
Summary:Climate models generally underestimate the pronounced warming in the sea surface temperature (SST) over the North Atlantic during the mid‐Pliocene that is suggested by proxy data. Here we investigate the influence of the North Atlantic cold SST bias, which is observed in many climate models, on the simulation of mid‐Pliocene surface climate in a series of simulations with the Kiel Climate Model. A surface freshwater‐flux correction is applied over the North Atlantic, which considerably improves simulation of North Atlantic Ocean circulation and SST under present‐day conditions. Using reconstructed mid‐Pliocene boundary conditions with closed Bering and Arctic Archipelago Straits, the corrected model depicts significantly reduced model‐proxy SST discrepancy in comparison to the uncorrected model. A key factor in reducing the discrepancy is the stronger and more sensitive Atlantic Meridional Overturning Circulation and poleward heat transport. We conclude that simulations of mid‐Pliocene surface climate over the North Atlantic can considerably benefit from alleviating model biases in this region.