hide
Free keywords:
-
Abstract:
Understanding the drivers and physical processes influencing Antarctic sea ice and being able to predict Antarctic sea ice is crucial for improving our current climate projections. We investigate Antarctic sea ice predictability using a high resolution global coupled ocean-sea ice model and evaluate it against observations. We explored the physical processes in the upper ocean underlying sea ice predictability and found that memory in the upper ocean resides largely within the winter water layer. The intensity and the depth to which ocean memory forms in a region is controlled by upper ocean vertical structure. Ocean memory responds to seasonal changes in the upper ocean, especially to the changes in stratification strength, driven mainly by sea ice processes. Our results present sea ice predictability as a signature of local ice-ocean interaction.<p> In our regional analysis, modelled and observed sea ice predictability diverge in most regions. Modelled sea ice produced higher predictability skills in summer and spring months, whereas in the observations, it is in winter. Assessing the spatial distribution of sea ice predictability skills, model and observation produced similar seasonal patterns, however, the predictability skill in the model was considerably higher than observation. Given, the sparsity of continuous ocean observation, especially under Antarctic sea ice, it is hard to pin point the differences in the representation of ocean conditions in these high latitudes between the model and observation. However, our findings indicate that better representation of upper ocean water column and its seasonal processes can improve Antarctic sea ice predictability skills.
