Abstract
Abstract. Sea ice in both polar regions is an important indicator of the expression of global climate change and its polar amplification. Consequently, broad interest exists on sea ice coverage, variability and long-term change. However, its predictability is complex and it depends strongly on different atmospheric and oceanic parameters. In order to provide insights into the potential development of a monthly/seasonal signal of sea ice evolution, we applied a robust statistical model based on different oceanic and atmospheric parameters to calculate an estimate of the September sea ice extent (SSIE) on a monthly timescale. Although previous statistical attempts of monthly/seasonal SSIE forecasts show a relatively reduced skill, when the trend is removed, we show here that the September sea ice extent has a high predictive skill, up to 4 months ahead, based on previous months' oceanic and atmospheric conditions. Our statistical model skillfully captures the interannual variability of the SSIE and could provide a valuable tool for identifying relevant regions and oceanic and atmospheric parameters that are important for the sea ice development in the Arctic and for detecting sensitive/critical regions in global coupled climate models with a focus on sea ice formation.
Highlights
Arctic sea ice plays an important role in modulating the global climate system by influencing the atmospheric and oceanic circulation in polar regions
Here we investigate the potential link between the Arctic September sea ice extent (SSIE) (Fetterer et al, 2016) and OHC, over the first 100 m (OT100) (Levitus et al, 2012; Boyer et al, 2013) and sea surface temperature (SST) (Huang et al, 2014) as longterm predictors (lags ∼ 4 years (AMO index) up to 2 months in advance; see Table 2 for a detailed description of all the lags used in the study)
The atmospheric circulation can substantially contribute to the skill of the sea ice predictions
Summary
Arctic sea ice plays an important role in modulating the global climate system by influencing the atmospheric and oceanic circulation in polar regions. This is particular true for the years with extreme low September sea ice concentrations (e.g., 2012 or 2007), with both the dynamical and the statistical models showing similar limitations (Stroeve et al, 2014, 2015; Schröder et al, 2014; Hamilton and Stroeve, 2016). In order to improve the monthly/seasonal prediction skill of the sea ice extent, one possibility would be to identify stable predictors (the correlation coefficient between the predictor and the predictand does not change in time) and to develop a statistical forecast model based on these predictors Following this idea, here we analyze the oceanic and atmospheric conditions associated with the SSIE in order to identify potential predictors based on a simple statistical methodology and place them in a longer temporal context.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
