Thanks for all your interests in our paper. I told Rob when I got his first email that I would like to discuss with anyone on stoat about our study. Since most scientific discussions about our paper have been happening on ASIF, I decide to start a discussion here. But what I will say only represents my own personal view and not that from the other 10 coauthors. I want to make this point very clear in the beginning.
I think our paper is not a study to play with those trends/interdecadal changes or regression procedures. We actually provide a different perspective to understand the warming process in the arctic in the past three-four decades.
In the paper, we didn't only focus on temporal variability. We focused on spatial changes of the large scale circulations in the Arctic.
Under CO2 forcing, the ensemble mean of all IPCC models ( these are almost the best models we can use, trust and learn from) gives us a very uniform rise of Z200 everywhere in the Arctic and a much less rise in the lower levels of the troposphere (Fig. 4 of our paper). This is because uniform CO2 warming( the greenhouse effect) will increase the depth of each atmospheric layer a little bit and thus the heights in the upper levels have the largest increase because of a cumulative effect of all small increases of layers below. So we can only see less increases of the heights in the lower levels but higher increases of heights at the higher levels.
In contrast , the observation show a very different structure with the most significant rise of geopotential height over Greenland at all levels from the surface to the upper troposphere. we call this type of change as a barotropic structure change.
I think my understanding of this difference is that the observed circulation pattern is a classical heat wave pattern that favors maximum warming in the lower boundary to melt sea ice and the build-up of the heat wave pattern over the Arctic in the past decades is due to some low-frequency variability of atmospheric Rossby wave originated from the tropics.
In other words, the wind changes due to some remote forcing in the past 30to 40 years pushed more air masses into the Arctic and then air there became more condensed and warmed. I think this is an important dynamical factor that caused so fast sea ice melting in the past decades.
It is still under debate whether this remote tropical forcing is internal or forced by the CO2 rise. What I learned from those IPCC models is that the observed tropical forcing pattern since 1979, especially in the recent decade is not the pattern favored by CO2 forcing. Please see our 2014 paper on this tropical linkage.
please see here about heat wave
http://scijinks.gov/heat/or here on wiki page: the formation part
https://en.wikipedia.org/wiki/Heat_wave however, the temperature warming pattern/structure favored by anthropogenic forcing in JJA in the Arctic is uniform everywhere with weaker intensity in the boundary layer. That is why, ( my personal view), most iPCC models cannot reproduce so fast sea melting as that observed if the models are only forced by Co2 forcing.
An analogy of this comparison is that anthropogenic forcing on the Arctic can be looked as turning on a furnace to warm a room. Internal forcing looks like putting a heated blanket on someone who feels cold. So this blanket ( as extra warming in the boundary layer in the real world) warms the guy faster.
In this paper, we just tried to use a model approach to remove this blanket and see how the room temperature increase melts sea ice. We also realized that some heating from the heated blanket to ice may also be due to the room temperature increase. In our attempt to do a final attribution analysis, we further removed that part and eventually got to that 40% contribution.