Search

Marine Isotope Stage (MIS) 5e (130–116 ka) represents a "laboratory" for evaluating climate model performance under warmer‐than‐present conditions. Climate model simulations for MIS 5e have previously failed to produce Southern Ocean (SO) sea‐surface temperatures (SSTs) and sea‐ice extent reconstructed from marine sediment core proxy records. Here we compare state of the art HadGEM3 and HadCM3 simulations of Peak MIS 5e SO summer SSTs and September sea‐ice concentrations with the latest marine sediment core proxy data. The model outputs and proxy records show the least consistency in the regions located near the present‐day SO gyre boundaries, implying the possibility that model simulations are currently unable to fully realize changes in gyre extent and position during MIS 5e. Including Heinrich 11 meltwater forcing in Peak MIS 5e climate simulations improves the likeness to proxy data but it is clear that longer (3–4 ka) run times are required to fully test the consistency between models and data. Plain Language Summary: Investigating past warm periods can provide us with an analog for how climate will respond to future warming. In this study we compare the latest model simulations of Southern Ocean (SO) sea‐surface temperatures and Antarctic winter sea‐ice extent from 130,000 years ago with data from marine sediment cores. The simulations and sediment core data show the least match in the areas near the boundaries of the present day SO gyres (large, circulating ocean currents), implying that possibly changes in the position and size of the gyres are not fully recreated in the computer simulations. The inclusion of ice sheet meltwater into the North Atlantic improves the comparison between the simulations and sediment core data but it is clear that longer run times are required to fully test their consistency. Key Points: Different areas of the Southern Ocean (SO) show different consistency between model simulations and proxy dataRegions along the edges of the modern day SO gyres show the least consistency between models and proxy dataInclusion of Heinrich 11 meltwater forcing improves the match between model simulations and proxy data [ABSTRACT FROM AUTHOR]

The latitudinal temperature gradient is a fundamental state parameter of the climate system tied to the dynamics of heat transport and radiative transfer. Thus, it is a primary target for temperature proxy reconstructions and global climate models. However, reconstructing the latitudinal temperature gradient in past climates remains challenging due to the scarcity of appropriate proxy records and large proxy-model disagreements. Here, we develop methods leveraging an extensive compilation of planktonic foraminifera d18O to reconstruct a continuous record of the latitudinal sea-surface temperature (SST) gradient over the last 95 million years (My). We find that latitudinal SST gradients ranged from 26.5 to 15.3 °C over a mean global SST range of 15.3 to 32.5 °C, with the highest gradients during the coldest intervals of time. From this relationship, we calculate a polar amplification factor (PAF; the ratio of change in >60° S SST to change in global mean SST) of 1.44 6 0.15. Our results are closer to model predictions than previous proxy-based estimates, primarily because d18O-based high-latitude SST estimates more closely track benthic temperatures, yielding higher gradients. The consistent covariance of d18O values in low- and high-latitude planktonic foraminifera and in benthic foraminifera, across numerous climate states, suggests a fundamental constraint on multiple aspects of the climate system, linking deep-sea temperatures, the latitudinal SST gradient, and global mean SSTs across large changes in atmospheric CO2, continental configuration, oceanic gateways, and the extent of continental ice sheets. This implies an important underlying, internally driven predictability of the climate system in vastly different background states. [ABSTRACT FROM AUTHOR]

SignificanceThe temperature difference between low and high latitudes is one measure of the efficiency of the global climate system in redistributing heat and is used to test the ability of models to represent the climate system through time. Here, we show that the latitudinal temperature gradient has exhibited a consistent inverse relationship with global mean sea-surface temperature for at least the past 95 million years. Our results help reduce conflicts between climate models and empirical estimates of temperature and argue for a fundamental consistency in the dynamics of heat transport and radiative transfer across vastly different background states.