Climate-related variability of isotopic records in a coastal Antarctic ice core

Abstract

Polar ice cores, both in Greenland and Antarctica, provide an improved understanding of past climatic variations through ⸹18O and ⸹D records, which are considered a proxy of surface temperature. This study focusses on the isotopic records obtained from four ice cores drilled in a coastal site of East Antarctica (GV7), characterized by a relatively high snow accumulation rate (237 mm of water equivalent). At this site different cores have been retrieved, one of which reached a depth of 250 m, while others reached shallow (5-12 m) or intermediate (55 m) depths, covering different time periods; this allows to calculate a stacked record for this site, in order to reduce the stratigraphic noise. The co-isotopic analysis allows the definition of the deuterium excess (d=⸹D-8*⸹18O), which is a proxy of climatic conditions (relative humidity, sea surface temperature, and wind speed) in the moisture source regions of precipitation. Since the isotopic records are not only related to temperature but also to other factors (precipitation intermittency, snow redistribution by winds and post-depositional effects), we compare the mean annual GV7 values of ⸹18O,⸹D, d and snow accumulation rate, all of which determined after an annual layer dating approach, to re-analysis data over the 1979-2012 period. For validating the proxy records, we used the ERA-Interim reanalysis data of temperature and precipitation, as well as the ECHAM5-wiso model data, which includes also the stable isotopes. We found that ERA-Interim over-estimates the precipitation data at GV7, which is already known in literature in coastal areas of Antarctica, while the ECHAM5-wiso under-estimates them. The simulated temperature data from ECHAM5-wiso were found in strong agreement with stable isotopes from GV7 ice cores, when using a 3-year moving average, except for one of them. Furthermore, we also compared the d-excess records with sea ice extent, finding a contrasting behavior in the most recent period. We also investigated the snow accumulation records in relation to simulated temperature, finding an agreement between the two variables from 1979 to 1996, while they appear not correlated afterward. The main climatic indices for the Southern Hemisphere, El Niño 3.4 and the Antarctic Oscillation, were also compared to stable isotopes and simulated temperatures. The overall model-data comparison in this study and for this area, shows a relatively good agreement, both in terms of absolute values and variability.