“…While the sign of the summer NAO does not look significantly different for 2070–2099 in comparison to past periods in CESM1‐LE (Figure b), the CMIP5 ensemble mean projects a slightly more positive summer NAO under future greenhouse gas forcing (Fettweis, Hanna, et al, ). However, the CMIP5 models are known to generally underestimate observed blocking in the high‐latitude North Atlantic and Europe (Anstey et al, ; Dunn‐Sigouin & Son, ; Masato et al, ; Hanna, Fettweis, & Hall, ), calling into question their ability to project future NAO changes and highlighting a need for better model representation of blocking and further study into the impacts of climate change on natural variability. An additional open question concerns the relative importance of different mechanisms for connecting high‐pressure conditions to extreme Greenland melt in different periods, as even for recent decades the mechanisms of warm air advection, cloud and radiative impacts, and large‐scale sinking and adiabatic warming have all been implicated (Ding et al, ; Fettweis et al, ; Hofer et al, ; Lim et al, ).…”