Antarctic sea ice response
Antarctic sea ice response in ensemble CMIP5 historical and ozone perturbation simulations
The net Southern Hemisphere sea ice area has a small statistically significant positive trend in recent decades in contrast to the decay in the Arctic sea ice cover. There have been a number of theories suggesting that it is driven by strengthening of SAM as a result of ozone depletion; increasing stratification of the ocean from sea ice melt and/or ice shelf melting, increased divergence of the ice pack or could be part of natural variability of the climate system. There have been several recent papers investigating the role of ozone depletion in individual models where they concluded ozone did not contribute to the increasing sea ice trend; other recent studies focus on the ice trends being driven by trends in the ice drift and trends in the winds.
The CMIP5 models provide a data set to investigate several of these theories; in particular the CSIRO Mk3.6 model has a 10 member historical ensemble in which the sea ice concentration has individual members that show variation in the trend of the net sea ice cover over the recent period (1976-2005). By studying the atmospheric drivers on the sea ice, air temperature, large scale circulation, surface, wind stress, the sea ice transport and sea ice divergence of individual ensemble members at monthly/seasonal timescale we hope to understand why the differences arise. We also investigate the changes in ocean temperature salinity and the ocean stratification in the individual members.
As part of CMIP5, many of the modelling groups undertook either ozone only historical runs or all forcing historical runs that omitted the recent ozone perturbations. CSIRO Mk3.6 undertook the latter, with a 10 member ensemble, and these runs are close to the present day signal so can readily be compared to the existing historical runs and observations. These data have been investigated to ascertain if the ozone signal in this model impacted on the ensemble spread of the results and how individual members of the ensemble respond compared to the all forcing case. Preliminary results showed that this ensemble also had members where the ice cover increased whilst others showed decrease in the net sea ice, suggesting a similar variation to the original ensemble. Further analysis will focus on whether the sea ice response in individual sectors of the Antarctic ice cover resembles the recent changes seen in the observations. The observational data sets to be used will be the satellite derived ice concentration available from 1979 and the meteorological reanalyses data and will build on existing studies of the co-authors.
A similar set of experiments has been just been completed for the ACCESS 1.0 and ACCESS 1.3 model with the ozone distribution at high southern latitudes remaining held at 1960 values until 2020. The ensemble size is smaller at only 3 members which is the same size as the existing all forcing historical ACCESS ensembles (the model is considerably more computational expensive to run). We hope that the ensemble size is sufficient to detect a signal, if it is present in either of the ACCESS model configurations. Analysis of these ensembles is about to be undertaken.