Climate and vegetation changes around the Atlantic Ocean resulting from changes in the meridional overturning circulation during deglaciation

Abstract

Abstract. The Bølling-Allerød (BA, starting ~ 14.5 ka BP) is one of the most pronounced abrupt warming periods recorded in ice and pollen proxies. The leading explanation of the cause of this warming is a sudden increase in the rate of deepwater formation in the North Atlantic Ocean and the resulting effect on the heat transport by the Atlantic Meridional Overturning Circulation (AMOC). In this study, we used the University of Victoria (UVic) Earth System-Climate Model (ESCM) to run simulations, in which a freshwater perturbation initiated a BA-like warming period. We found that under present climate conditions, the AMOC intensified when freshwater was added to the Southern Ocean. However, under Heinrich event 1 (HE1, ~ 16 ka BP) climate conditions, the AMOC only intensified when freshwater was extracted from the North Atlantic Ocean, possibly corresponding to an increase in evaporation or a decrease in precipitation in this region. The intensified AMOC led to a warming in the North Atlantic Ocean and a cooling in the South Atlantic Ocean, resembling the bipolar seesaw pattern typical of the last glacial period. In addition to the physical response, we also studied the simulated vegetation response around the Atlantic Ocean region. Corresponding with the bipolar seesaw hypothesis, the rainbelt associated with the Intertropical Convergence Zone (ITCZ) shifted northward and affected the vegetation pattern in the tropics. The most sensitive vegetation area was found in tropical Africa, where grass cover increased and tree cover decreased under dry climate conditions. An equal but opposite response to the collapse and recovery of the AMOC implied that the change in vegetation cover was transient and robust to an abrupt climate change such as during the BA period, which is also supported by paleovegetation data. The results are in agreement with paleovegetation records from Western tropical Africa, which also show a reduction in forest cover during this time period. Further agreement between data and model results was found for the uplands of North America and Southern Europe, where grassland along with warm and dry climates were simulated. However, our model simulated vegetation changes in South and North America that were much smaller than reconstructed. Along the west and east coast of North America we simulated drier vegetation than the pollen records suggest.