Combining ice core records and ice sheet models to explore the evolution of the East Antarctic Ice sheet during the Last Interglacial period

This study evaluates the influence of plausible changes in East Antarctic Ice sheet (EAIS) thickness and the subsequent glacio-isostatic response as a contributor to the Antarctic warming indicated by ice core records during the Last Interglacial period (LIG). These higher temperatures have been estimated primarily using the difference in the δD peak (on average~15‰) in these LIG records relative to records for the Present Interglacial (PIG). Using a preliminary exploratory modelling study, it is shown that introducing a relatively moderate reduction in the amount of thickening of the EAIS over the LIG period introduces a significant increase (up to 8‰) in the predicted elevation-driven only δD signal at the central Antarctic Ice sheet (AIS) ice core sites compared to the PIG. A sensitivity test in response to a large prescribed retreat of marine-based ice in the Wilkes and Aurora subglacial basins (equivalent to ~7m of global mean sea-level rise) results in a distinct elevation signal that is resolvable within the ice core stable isotope records at three sites (Taylor Dome, TALDICE and EPICA Dome C). These findings have two main implications. First, EAIS elevation's only effects could account for a significant fraction of the LIG warming interpreted from ice core records. This result highlights the need for an improved estimate to be made of the uncertainty and size of this elevation-driven δD signal which contributes to this LIG warming and that these effects need to be deconvolved prior to attempting to extract a climatic-only signal from the stable isotope data. Second, a fingerprint of significant retreat of ice in the Wilkes and Aurora basins should be detectable from ice core δD records proximal to these basins and therefore used to constrain their contribution to elevated LIG sea levels, after accounting for ice sheet–climate interactions not considered in our approach.