Antarctic ice mass variations from 1979 to 2017 driven by anomalous precipitation accumulation

Global mean sea level (GMSL) is rising, primarily due to contributions from melting mountain glaciers and the Greenland Ice Sheet (GrIS). The Antarctic Ice Sheet (AIS) has also contributed, with the rate of ice mass loss increasing in recent decades. The AIS ice mass balance is determined by ice discharge and surface mass balance (SMB), the latter primarily influenced by precipitation. While previous research emphasized the role of ice discharge in long-term AIS mass variations, this study revisits this assumption by examining the contribution of precipitation to Antarctic ice mass changes over recent decades. Understanding the relative contributions of precipitation and ice discharge is crucial for projecting future ice mass loss and its impact on sea level rise. Anomalous precipitation has been shown to influence Antarctic ice mass loss estimates, highlighting the need for a comprehensive understanding of its role in future projections. This study aims to re-evaluate Antarctic ice mass changes and quantify the contribution of precipitation to these changes.

Existing studies have shown an increase in the rate of GMSL rise in recent decades, largely attributed to water inflow from mountain glaciers and the GrIS. The AIS's contribution to GMSL has been estimated to be about 0.3 mm/year, roughly half that of the GrIS, although recent studies (IMBIE2) suggest a significant acceleration of AIS ice mass loss in the last decade. Previous research has largely focused on ice discharge as the primary driver of long-term AIS mass balance changes, assuming minor contributions from precipitation variations. However, recent findings based on GRACE gravity data and SMB reanalysis models indicate that precipitation decrease can account for a significant portion of ice mass loss acceleration, particularly in the Amundsen Sea Embayment. These studies emphasized the importance of considering precipitation effects on AIS mass balance for a better understanding of multi-decadal and longer-term changes.

This study uses the state-of-the-art ERA5 reanalysis data to examine AIS SMB variations from 1979 to 2017. The researchers analyzed accumulated precipitation (time-integrated precipitation), which is approximately equivalent to SMB, to capture long-term variations. A linear trend was removed from the accumulated precipitation time series to focus on inter-annual and longer variations. Principal component analysis (PCA) was employed to identify major patterns in the detrended SMB (ΔSMB) field. Rotated EOF (REOF) analysis was further utilized to disentangle the effects of different atmospheric circulation modes (Southern Annular Mode (SAM) and El Niño-Southern Oscillation (ENSO)) on SMB variations. Regression analysis was also used to examine the relationship between ΔSMB and ΔSAM. To validate the ERA5 model results, the researchers compared model-based SMB estimates with satellite gravimetry (GRACE) observations. Corrections were applied to the GRACE data to account for atmospheric pressure errors and ice discharge acceleration. Finally, the researchers combined IMBIE2 estimates of AIS mass change with ERA5 SMB estimates to analyze the contribution of SMB to the observed ice mass loss acceleration.

The study found significant inter-annual and multi-decadal variations in AIS SMB, despite near-zero long-term trends in precipitation rates. Multi-decadal SMB variations were strongly linked to the SAM, exhibiting a distinct bi-polar pattern: negative acceleration in the Pacific sector and positive acceleration in the Atlantic and Indian sectors. PCA and REOF analysis confirmed the dominant influence of SAM on SMB variations. The correlation coefficient between the first REOF mode PC and ΔSAM was 0.89, indicating a strong relationship. Model predictions of SMB variations were validated by comparing ERA5-derived ΔSMB with GRACE observations after correcting for barometric pressure errors and ice discharge. After correcting for SMB contributions, the study found a steady acceleration of ice discharge of -8.7 ± 0.3 Gton/year² from 1992 to 2017. The apparent abrupt change in ice mass loss around 2007 was attributed to variations in SMB, primarily in West Antarctica. Approximately 27% of the increase in Antarctic ice mass loss from 1992-2006 to 2007-2017 (147 Gton/year) was attributed to SMB variation, with a more significant impact (41%) observed in West Antarctica.

The findings highlight the importance of considering SMB variations, particularly those driven by SAM, when assessing AIS mass balance and projecting future sea-level rise. The study demonstrates that the previously reported abrupt increase in Antarctic ice mass loss around 2007 is not solely due to changes in ice dynamics (ice discharge), but also significantly influenced by SMB variations. The bi-polar pattern in SMB changes driven by SAM has important implications for regional ice mass balance and its contribution to global sea level. This study challenges the common assumption that long-term AIS mass variations are mainly controlled by ice discharge, revealing the significant and previously underestimated role of SMB, especially in the context of climate variability.

This research demonstrates the significant impact of SMB variations, particularly those linked to the SAM, on AIS mass balance. The study's findings challenge the traditional focus on ice discharge as the primary driver of long-term AIS mass changes. The observed abrupt increase in AIS mass loss around 2007 is primarily attributed to SMB variations, not a sudden change in ice dynamics. Further research should focus on improving the accuracy of SMB estimations and examining the interactions between climate variability, SMB, and ice dynamics to enhance the accuracy of future sea-level rise projections. The bi-polar pattern of SMB acceleration, controlled by SAM, also warrants further investigation.

The study relies heavily on reanalysis data (ERA5), which has inherent uncertainties. While the authors validated their SMB estimates using GRACE data, there are still uncertainties associated with GRACE measurements (e.g., atmospheric pressure corrections). The study focuses on the impact of SAM, but other climate modes might also influence SMB variations. Furthermore, the simplified approach of subtracting ΔD from ΔM in the GRACE data analysis introduces some degree of error. The study’s approach primarily focused on the accumulated precipitation and therefore, its analysis might not fully capture the impacts of higher-frequency changes in precipitation rates.