Attribution And Future Projections Of Long-term Changes In The Antarctic Circumpolar Wav

Antarctic Circumpolar Wave（ACW）is one of the most important members in the Southern Hemisphere（SH）climate system founded in recent decades.It has substantial impacts on the SH and global climate.As an interannual large-scale ocean–atmosphere coupling system,ACW appears as the large-scale wave of mid-to-high latitude variability anomalies propagating eastwards around the Antarctic.In this research,the latest long period reanalysis dataset is used to reconstruct the 180-year ACW series after extracting from sea level pressure（SLP）and sea surface temperature（SST）fields during 1836–2015.The ACW is found having long-term variation in the nearly two centuries.Its relationship with associated large-scale climate indices is also discussed.Using 24 CMIP5 and CMIP6 model historical experiment outputs,we qualitatively and quantitively evaluate their performance in simulating ACW SLP and SST signals during its most active period and find the best one.Afterwards CMIP6 model simulations under different external forcing are used to find out the response of ACW,then detect and attribute the ACW variation in the late 20^th century.Finally,we give out the long-term climate prediction of ACW based on 11 CMIP6 model SSP2-4.5 and SSP5-8.5 experiment outputs.Major conclusions are as follows:1.The ACW has significant long-term variation during the past 180 years.The ACW SLP and SST signals can be linearly decomposed into the ACW-2 and ACW-3 components,with the ACW-2 pattern dominating.The SLP signal is highly corelated with the SST signal and about2 months ahead of the last.The SLP signal has an interdecadal phase shift in the 1950s.It may be related to the interdecadal variation of the SAM and PSA.The interdecadal shift of SST signal appears in the 1970s,which may be attributed to the change of ENSO.The 1980s-1990s is the most pronounced period of ACW.However,there is no evidence of change in zonal distribution of ACW amplitude.The ACW is most significant in the South Pacific,while it is most fuzzy in the southern Indian Ocean.2.When take all CMIP5 and CMIP6 models into consideration,models simulate monthly mean SST better than SLP in the mid-to-high latitudes of SH.The CMIP6 models show a slightly better performance than the CMIP5 models in reproducing ACW SLP and SST signals.GISS-E2-1-H is the best model simulating ACW signals among all 24 investigated models.The main problem of model simulation is a low amplitude of SST signal.This may come from the poor performance in reproducing the leading ACW-2 component.3.Both ACW SLP and SST signals’strength show a significant increasing trend in the last half of 20^th century.The increase of SLP signal’s strength is mainly attributed to human factors.The increase of SST signal’s strength can partly be attributed to human factors.4.In the late 21^stcentury SLP filed shows a pattern of decreasing high-latitudes and increasing mid-latitudes in both scenarios.The SST field shows a pattern of warming high-latitudes and warmer mid-latitudes in both scenarios.SSP5-8.5 experiences more intense change than SSP2-4.5 scenario.The zonal distributions of simulated ACW SLP and SST signals vary from models and scenarios.The amplitude of multi-model-mean result decreases in both scenarios.A zonal distribution change of multi-model-mean result is seen in SSP5-8.5 scenario.The ACW SLP and SST indices also show large variation among single-model predictions.The multi-model-mean result experience larger amplitude and pronounced wave characteristics in both scenarios.