Springtime winds drive Ross Sea ice variability and change in the following autumn

Marika M Holland,Marilyn Raphael,Laura Landrum,Sharon Stammerjohn

doi:10.1038/s41467-017-00820-0

Abstract

Autumn sea ice trends in the western Ross Sea dominate increases in Antarctic sea ice and are outside the range simulated by climate models. Here we use a number of independent data sets to show that variability in western Ross Sea autumn ice conditions is largely driven by springtime zonal winds in the high latitude South Pacific, with a lead-time of 5 months. Enhanced zonal winds dynamically thin the ice, allowing an earlier melt out, enhanced solar absorption, and reduced ice cover the next autumn. This seasonal lag relationship has implications for sea ice prediction. Given a weakening trend in springtime zonal winds, this lagged relationship can also explain an important fraction of the observed sea ice increase. An analysis of climate models indicates that they simulate weaker relationships and wind trends than observed. This contributes to weak western Ross Sea ice trends in climate model simulations.

Highlights

Autumn sea ice trends in the western Ross Sea dominate increases in Antarctic sea ice and are outside the range simulated by climate models
The western Ross Sea ice trends do not appear attributable to observed changes in many modes of variability[6], including the significant trends in the austral summer Southern Annular Mode[7], which are associated with stratospheric ozone loss
October zonal wind anomalies in the high latitude South Pacific drive a substantial fraction of the western Ross Sea ice variability in the following autumn

Summary

Introduction

Autumn sea ice trends in the western Ross Sea dominate increases in Antarctic sea ice and are outside the range simulated by climate models. While previous work has indicated that the observed increases in the total Antarctic sea ice extent are consistent with natural variability simulated by climate models[4], this is not the case if the regional and seasonal nature of the trends is considered. We use a number of independent data sets to assess factors that contribute to variability and change in western Ross Sea ice cover during autumn This includes the analysis of mechanisms driving a seasonal memory of the sea ice and the possible consequences for observed sea ice trends in the western Ross Sea. We show that, on interannual timescales, wind variability in the previous October is a strong predictor of ice anomalies in March to May, with implications for seasonal ice predictability. Climate models simulate weaker relationships and wind trends than observed, which

Methods

Results

Conclusion