Long-term historical climate and surface mass balance over the Antarctic Ice Sheet

Abstract

Although the climate of the Antarctic Ice Sheet has been subject to simulations for many years, they are generally limited to the past few decades, rendering a good assessment of the past climate including the processes linked to the variability of the latter difficult. In this study, a recently released back extension of the ERA5 reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ECMWF) was used for forcing the polar-oriented regional climate model MARv3.12 over the Antarctic Ice Sheet from 1950 to 2020 using a 35 km resolution. Regarding near-surface pressure and temperature (SAT), MARv3.12 displays good performance for the back-extension period (1950–1978) and properly reproduces the SAT variability across the AIS. The model slightly underestimates near-surface temperature along the coastline and tends to overestimate values on the Antarctic Plateau. Likewise, MARv3.12 forced by ERA5 appears to accurately portray surface mass balance (SMB) for both 1950–1978 (r = 0.85) and 1979–2020 (r = 0.85). Mean biases (-8.5 kg m-2 yr-1and -37.8 kg m-2 yr-1) are to a certain extent the result of drifting snow processes as well as atmospheric sublimation not being included in model calculations. An assessment of the evolution of the climate reveals an overall increase in SAT, SMB, and surface melt (SM) over time. Especially high temperature increases (> 2°C between the 1951–1980 and 1981–2010 mean) are found in East Antarctica. In summer, across the area Ross ice shelf, both SAT and SMB exhibit a decrease. Interestingly, MARv3.12 forced by ERA5 produces a pronounced surge in SAT and SMB in the late 1970s, notably during austral winter (JJA). Since the same pattern is displayed by the reanalysis itself, this jump may be artificial and a result of the assimilation of stratospheric ozone measurements into the ERA5 dataset around 1979. This hypothesis is supported by the fact that the abnormal rise in temperature does not appear to be present in observation-based climate records. Moreover, in recent studies, a switch in the relationship between the SAT and the Southern Annular Mode (SAM) was identified in the late 1970s, a feature also replicated in this study using MARv3.12 forced by ERA5. It remains unclear if there is a connection between these two events occurring around 1980, however, in future studies MAR could be forced by different reanalyzes also spanning over a longer period or the results presented in this study could be directly compared to ice core measurements. Either way, this study demonstrates the usability of the ERA5 reanalysis for early climate simulations and identifies areas where further research is necessary for assessing the evolution of the Antarctic climate.