ABSTRACT

Rapid changes in the Arctic have been observed and are expected to be amplified in future. However, the influence of Arctic sea ice decline on global climate (including the Pacific) remains under debate. First, I will focus on a modeling analysis of the ocean-atmosphere responses to a near-term Arctic sea ice loss (corresponding to the period 2035-2055 under the RCP8.5 scenario). With the coupled model IPSL-CM5A2, two different methods to melt Arctic sea ice are implemented to estimate the robustness of the protocol. I will show that modifying the albedo or changing the thermal conductivity of the Arctic sea ice show similar pattern responses but with different magnitude. It is found that, at the 90% significance level, Arctic sea ice loss leads to large-scale circulation responses, such as a negative NAO-like associated with southward shift of the westerlies; expansion of the North Atlantic current; a weaker AMOC; a southern shift of the ITCZ; tropical Pacific cooling and a positive SAM. Then, I will present an observational analysis of the direct atmospheric response to sea ice variability for the period 1979-2016. By performing lagged regressions with a dominant Arctic sea ice cover pattern, no significant large-scale field response is detected in October, November, or April. A negative NAO-like pattern via a downward stratospheric pathway is established from December to March. Lagged multiple regressions including concomitant Siberian snow cover and Pacific SST fluctuations demonstrate the important role of snow for the stratospheric forcing pathway, while a tropospheric pathway is attributed to Arctic sea ice cover. These studies emphasize the need for coordinated experiments to extract the role of Arctic sea ice on global climate.