Cooling of Eurasian Winters Over Last 25 Years Not Likely to Be Due to Loss of Arctic Sea-Ice

Australia: The Land Where Time Began

Cooling of Eurasian Winters Over Last 25 Years Not Likely to Be Due to Loss of Arctic Sea-Ice

Over the past 25 winters surface air temperatures above central Eurasia decreased at a time of strongly increasing anthropogenic forcing and Arctic Amplification. Loss of sea ice in the Barents-Kara Sea has been suggested as being related to this cooling. In this paper McCusker et al. used more than 600 years of atmosphere-only global climate model simulations to isolate the effect of the loss of Arctic sea ice, which was complemented with a 50-member ensemble of atmosphere-ocean global climate model simulations while allowing for forcing changes, both anthropogenic and natural, and internal variability. No evidence was found in their atmosphere-only simulations of the loss of Arctic sea ice having impacted Eurasian surface temperatures. It was found in their atmosphere-ocean simulations there only 1 simulation with Eurasian cooling that was of the observed magnitude, though loss of sea-ice was not involved, directly or indirectly. In this simulation it was found that the cooling is due to a certain pattern of circulation which combined high pressure above the Barents-Kara Sea and a downstream trough. It was concluded that an internally generated pattern circulation that has been ensconced over, and nearby, the Barents-Kara Sea since 1980 was probably responsible for the cooling that has been observed over central Asia. Knowledge of high-latitude variability of climate and change has been improved by this study, with implications for an understanding of impacts in high northern latitude systems.

Surface air temperature (Hansen et al., 2010) (SAT) over central Eurasia (CEUR) in winter (December to February) has cooled over the past 25 years, while generally elsewhere in the Northern Hemisphere warming has prevailed, in what McCusker et al. suggest is likely to be a response to increasing anthropogenic radiative forcing (Myhre, G. et al., 2013). This cooling is seen in the difference between the surface air temperature over central Eurasia and the Northern Hemisphere. There was a decrease by more than 0.3 million square kilometres in the Arctic area of sea ice cover (Comiso, J. C., 2015) in the adjacent Barents-Kara Sea (BKS) over the same period. It is suggested by a small but significant correlation between interannual fluctuations in de-trended Barents-Kara Sea surface air temperature from 1979 to 2012 (r = 0.34 and ρ = 0.05) that the long-term central Eurasian cooling could be approximately ‘inferred’ from the Barents-Kara Sea loss of sea ice. This suggestion is supported by a regression-based procedure. This is, however, a statistical calculation that assumes a relationship between long-term trends as discussed below.

There are several atmospheric circulation pathways theoretically linking near-surface warming in the Arctic to cold mid-latitude winter weather on intraseasonal to interannual timescales (Cohen, J. et al., 2015). Cold winters in central Eurasia are correlated with increases in Barents-Kara Sea surface temperatures in observations (Outten, S. & Esau, I.A., 2012; Kug, J.-S. et al., 2015). Changes in the cover of sea ice could be the instigator of atmospheric circulation changes that cause outbreaks of continental cold air (Overland & Wang, 2010) and therefore cold winters in central Eurasia because it is believed amplification of near-surface temperature arises primarily from the loss of sea ice (Screen & Simmonds, 2010). Cold central Eurasian surface air temperature corresponds to low Barents-Kara Sea sea ice concentrations, that have been observed, is consistent with the above calculation (Inoue, Hori & Takaya, 2012; Overland, Wood & Wang, 2011; Francis et al., 2009; Tang et al., 2013), though this does not necessarily imply that changes in the concentration of BKS sea ice cause changes in the BKS or CEUR SAT, thus the need for modelling experiments that are targeted.

The results that have been obtained by modelling studies that attempted to ascertain if loss of BKS sea ice causes decreased CEUR SAT in winter have so far been mixed. It has been found by modelling studies that prescribe patterns of low and high concentrations of sea ice (SIC) over the BKS have found that patterns of low SIC are conducive to outbreaks of cold CEUR SAT on daily to interannual timescales (Honda, Inoue & Yamane, 2009; Mori et al., 2014; Kim et al., 2014; Petoukhov & Semenov, 2010). The magnitude of the central Eurasian cooling obtained by this is typically weaker than by observations (Mori et al., 2014; Kim et al., 2014; Petoukhov & Semenov, 2010), and in one case the central Eurasian cooling is apparent only in a subset of ensemble members (Honda, Inoue & Yamane, 2009). Therefore, though it is suggested by modelling studies that are based on cases of low and high sea ice concentration, that sea ice loss in the Barents-Kara Sea may decrease the central Eurasian surface air temperature, whether or not the effect is of large enough magnitude to be detected by observations over the influences of internal variability and/or warming that is externally forced remains an open question.

In this paper McCusker et al. address the distinct though related question of whether the declining trend in the Barents-Kara Sea sea ice has led to the central Eurasian cooling trend on a multi-decadal timescale that is observed, motivated by these studies and the interannual relationship that has been observed between the sea ice in the Barents-Kara Sea Central Eurasian surface air temperature that suggests there is a long-term connection. Simulations by an atmosphere-only global climate model (AGCM) that allow the isolation of the impact of the loss of Arctic sea ice in the absence of any direct forcing of central Eurasian surface air temperature. These AGCM simulations also allow for the explicit determination of cause and effect in a way that the statistical method that was discussed previously does not.

Sources & Further reading