Ocean, Climate, Paleoclimate and Geoscience

This essay based on my master thesis work is still under drafting. Please click on the title to download a full version of my thesis.

Abstract

The global mean ocean temperature (MOT) is an important measure for the state of Earth climate as it reflects the global energy balance. The potential of using noble gases as MOT proxies was seen due to their unique properties: they are chemically inert and their solubility is temperature-dependent. The solubility of noble gases in seawater decreases with increasing ocean temperature, and thus a lower MOT should result in a lower atmospheric inventory. This change can be measured in the noble gas mixing ratios from air bubbles trapped in ice cores.

This thesis aims to examine with the climate modelling approach the relation between the atmospheric Krypton mixing ratio anomaly (δKratm) and the MOT anomaly (∆MOT) under various boundary conditions. Sensitivity tests were per- formed with the Bern3D 2.0 Earth system model of intermediate complexity, which allows simulations to run on multi-millennial time-scale. Different boundary condi- tions were achieved by varying atmospheric CO2 concentration, diapycnal diffusiv- ity and wind stress. A new 5-box model was developed to further understand some results from the Bern3D model in a highly simplified setting.

Simulation results show that two factors, the sea ice and the state of the At- lantic Meridional Overturning Circulation (AMOC), have influence on noble gases responses to MOT change via air-sea gas exchange. Depending on the dominant factor in each simulation, δKratm’s response to MOT change can be divided into three regimes.

In simulations with ∆MOT>1.3°C, very little or zero sea ice exists in the Southern Ocean, resulting in a global mean Kr saturation ratio of higher than 98.5%. δKratm holds a relation with the MOT signal as directly calculated from the temperature-dependent solubility equation.

In the second regime where ∆MOT<1.3°C, the impact of sea ice becomes non- negligible on the δKratm/∆MOT relation. By blocking surface air-sea gas exchange, sea ice over the Antarctic Bottom Water (AABW) formation region can result in undersaturation in the deep ocean. This decouples δKratm from deep ocean temperature, and hence from the MOT signal. The global mean Kr saturation ratio decreases proportionally as sea ice expands over the Southern Ocean. Once the Southern Ocean is fully covered by sea ice, global mean Kr saturation ration remains stable even when sea ice expands further in other regions.

Sensitivity tests in both the Bern3D 2.0 model and the new 5-box model show that δKratm responds non-linearly to ice leak change. This justifies the omission of sea ice when converting ice-core measured δKratm to MOT anomalies in current research. Although, this could lead this could lead to a constant offset of 0.1‰ to 0.2‰ in the expected δKratm at a certain ∆MOT.

The third regime appears when AMOC is collapsed with wind stress lower than 70% of the control simulation. With sea ice over the Southern Ocean, a tropical- centralised surface heat distribution under weak wind stress leads to the result that only warm temperature signals are recorded by δKratm. As a result, δKratm is further decoupled from the MOT signal.

While proxy studies suggest that a collapsed AMOC existed in the past, the Bern3D 2.0 model was not able to reproduce it in LGM conditions, neither has there been strong disagreement between the current noble gas reconstructed MOT records and other temperature proxy reconstructions. Overall, noble gases can be used as reliable proxies for the past 800 kyr MOT reconstructions, although sea ice may cause small deficit by partially blocking the air-sea gas exchange and generating undersaturation. Rare situations are discussed when proxy reliability is compromised by shutdown of AMOC.