Prediction of Arctic Temperature and Sea Ice Using a High-Resolution Coupled Model

Under global warming, surface air temperature has risen rapidly and sea ice has decreased markedly in the Arctic. These drastic climate changes have brought about various severe impacts on the vulnerable environment and ecosystem there. Thus, accurate prediction of Arctic climate becomes more important than before. Here we examine the seasonal to interannual predictive skills of 2-m air temperature (2-m T) and sea ice cover (SIC) over the Arctic region (70 degrees-90 degrees N) during 1980-2014 with a high-resolution global coupled model called the Met Office Decadal Prediction System, version 3 (DePreSys3). The model captures well both the climatology and interannual variability of the Arctic 2-m T and SIC. Moreover, the anomaly correlation coefficient of Arctic-averaged 2-m T and SIC shows statistically significant skills at lead times up to 16 months. This is mainly due to the contribution of strong decadal trends. In addition, it is found that the peak warming trend of Arctic 2-m T lags the maximum decrease trend of SIC by 1 month, in association with the heat flux forcing from the ocean surface to lower atmosphere. While the predictive skill is generally much lower for the detrended variations, we find a close relationship between the tropical Pacific El Nino-Southern Oscillation and the Arctic detrended 2-m T anomalies. This indicates potential seasonal to interannual predictability of the Arctic natural variations.