Home > Research > Thesis and Habilitation thesis > Soutenances récentes et à venir > Regional variability of the polar snow densification during climatic transitions


OSUG - Terre Univers Environnement

Regional variability of the polar snow densification during climatic transitions

Soutenance de thèse de Camille Bréant, 17 novembre 2017

by Brice Boudevillain - 17 November 2017

The phasing between increases in temperature and greenhouse gas concentrations during large climatic variations in the past is classically estimated using analyses in polar ice cores, in the ice phase for the temperature and in the gas phase (trapped air bubbles) for the concentration of greenhouse gases. This phasing is still insufficiently constrained and solving this problem requires a better understanding of the mechanical process of snow to ice metamorphism near to the top of the ice sheet (i.e. the firn, about 100 m deep). In the absence of melting, the transformation of snow (a material with open porosity in contact with the atmosphere) into ice (a material containing isolated bubbles) occurs progressively as a response to temperature gradients near the surface, and the weight of overlying snow in deeper layers. Depending on temperature and precipitation conditions, this process occurs in a few decades to several millennia and a 100 meters depth range. It controls the age difference between the ice and the entrapped gases. Predicting the gas trapping depth is a major issue in paleoclimatology, especially in order to understand the phasing between temperature changes and changes in greenhouse gas concentrations.

A thermo-mechanical model of snow densification has been developed at LGGE, it includes the main mechanical processes, the thermal properties of ice, and gas trapping criteria. The model performances can be tested and improved using experimental studies of modern firns (density, open/closed porosity ratio, etc). For firnification under ancient climates, measurements of isotopes of inert gases (δ15N et δ40Ar) in the air trapped in ice cores provide direct informations about past variations of firn structure (e.g. diffusive zone thickness). Large differences between firn densification model outputs and gas isotopic data are obtained in Antarctica, and imply a large uncertainty on past climatic reconstructions. Understanding this discrepancy is a major issue in paleoclimatology.

As part of this thesis work, we took into account the effects of the temperature dependence of activation energies and impurities (dust) on the firn densification speed. It allowed to reconcile the model results with available data. The modified model results show an overall consistency with measured density profiles of present-day polar firns and isotopes of inert gases over deglaciations (also called terminations). We also analyze new high resolution measurements of δ15N and δ40Ar over Terminations 2 (129-138 ka) and 3 (243-251 ka) on the Dome C and Vostok ice cores. We have shown that the different evolutions of δ15N between different sites and different deglaciations are largely explained by differences in accumulation rates that control the snow/ice transition depth. We also showed that the use of air isotopes was an important complement to the use of water isotopes to constrain local climatic dynamics in eastern Antarctica during deglaciations.
In the course of this thesis, we showed that both the local temperature and accumulation conditions generally reconstructed from water isotopes records in Antarctica and the concentration of some chemical species are key parameters for the firn densification. In order to better understand these processes, we have explored how isotopic composition of water and chemical composition are linked in Antarctica during deglaciations but also on a daily scale thanks to a field mission that I carried out at the Antarctic base of Dumont d’Urville.