For most of the year arctic clouds are a positive feedback.
Determining when clouds are a negative feedback isn't simple. There are multiple cloud types varying in albedo, base height temperature, condensation nuclei radius, etc; there are variations in solar angle by latitude; and differences in the surface beneath the clouds. All of these can effect the shortwave/longwave radiation budget.
Most years the CAB surface is insulated from the ocean by a moderate layer of ice. This allows the SSTs to rise above the upper layer ocean temp of -2.4 C. The UNI Bremen N80 graphs show this as the temps are constrained by the melting point of snow and ice at the surface (freshwater) at approximately 0C.
This year we've seen widespread portions of the CAB have lower sea ice concentration. To me, this means that more of the ocean - at -2.4C - is exposed to the atmosphere. Rather than constraining the SSTs to 0 C, we're constraining it to a temperature closer to that of the ocean at -2.4C.
With air temps constrained below the melting point of snow, we should expect to see little if any surface melt, but increased bottom melt.
In other words, I don't believe the UNI Bremen N80 temp graph is telling us that 2013 is colder than previous years - it's telling us a different story - one that we haven't heard before. And I can't reconcile UNI Bremen with this: