The impact of winter ice retreat on Antarctic winter sea-ice extent and links to the atmospheric meridional circulation
15 years of weekly Antarctic ice extent and 10 m winds from numerical meteorological reanalyses have been used to test the hypothesis that the Antarctic final winter ice extent (FWE) is brought about by ice retreats as much as by advances, and that both are strongly affected by the meridional (north-south) component of the atmospheric circulation. This hypothesis is found to be correct and it is shown that extensive FWE fails to occur when total winter retreat is anomalously large. This is the case even in the coldest Antarctic regions. Retreats reduce the time available for advance, notably when they are substantial, with the ice cover taking up to several weeks to recover. Systematic changes in the meridional winds between retreats and advance are also detected in all regions; retreats are consistently associated with northerly winds supporting ice compaction and ice drift. The results agree with Antarctic case studies. Close similarities are found between several Antarctic and sub-Arctic regions in terms of the prevalence of retreat in some winters. its impact on the FWE and its relationship to the meridional atmospheric circulation. The study also reveals a more complex picture of the atmospheric circulation during sub-monthly ice retreats and advances. In particular, retreats in some Pacific regions are, on average, associated with anomalous anticyclonic circulations. This helps to explain why evidence of strong cyclone-ice extent relationships has not been found previously. A meteorological explanation is also sought for total retreat in winter being small in a winter of limited ice extent in the Bellingshausen Sea despite this being the mildest Antarctic region. In such cases, limited winter ice extent is attended by reduced advance caused by ice compaction due to northerly winds and waves and also anomalously high air temperatures. Reduced advance then limits opportunities for retreat to take place compared with winters with more advance and sea ice reaching lower latitudes. Overall. the results point to sub-monthly ice-atmospheric circulation interactions largely determining the winter ice extent throughout much of the Antarctic.