There is a new article out on those pesky Arctic clouds that climate models and Earth Energy Budget estimates don't even come close to getting right. This is the Abstract:
This study demonstrates that absorbed solar radiation (ASR) at the top of the atmosphere in early summer (May–July) plays a precursory role in determining the Arctic sea ice concentration (SIC) in late summer (August–October). The monthly ASR anomalies are obtained over the Arctic Ocean (65°N–90°N) from the Clouds and the Earth's Radiant Energy System during 2000–2013. The ASR changes primarily with cloud variation. We found that the ASR anomaly in early summer is significantly correlated with the SIC anomaly in late summer (correlation coefficient, r ≈ −0.8 with a lag of 1 to 4 months). The region exhibiting high (low) ASR anomalies and low (high) SIC anomalies varies yearly. The possible reason is that the solar heat input to ice is most effectively affected by the cloud shielding effect under the maximum TOA solar radiation in June and amplified by the ice-albedo feedback. This intimate delayed ASR-SIC relationship is not represented in most of current climate models. Rather, the models tend to over-emphasize internal sea ice processes in summer.
Every since I noticed the Earth Energy Budget screw up I have focused on mixed phase clouds, the "atmospheric window" and the very simple approximation of incident Solar radiation. Liquid topped mixed phase clouds produce a radiant "ground plane" of sorts that really should be treated as a different "surface". The models and the "experts" mention that sea ice melt would increase the albedo of the polar oceans allowing greater ocean heat uptake. That is true, but direct solar isn't really the issue in the Arctic due to the low solar angle of incidence. Thanks to the atmospheric lens effect and the liquid water surface of the mixed phase clouds, how much the lensing, mixed phase cloud area and sea ice area all interact to increase or decrease Arctic ocean heat content is one hell of a nifty puzzle. Older papers I have perused indicate around 18Wm-2 of uncertainty which is about what I estimated from the older K&T energy budgets. Soon to be if not already Dr. Barrett's thesis indicated that "global" liquid topped mixed phase clouds cover around 7.8 percent of the surface at any given time. For a quick estimate, that could produce around 15 Wm-2 +/- 10 of uncertainty which as you can see is a pretty large WAG. It could be that mixed phase clouds account for the majority of the model error. Don't know of course, but it is in the ballpark.
The models wouldn't just rely on the crude estimates tossed around by the online groupies. They would attempt to use actual incident radiation from both the sun and the clouds, provided they get the clouds right. Mixed phase clouds, they obviously don't, but it looks like the other types of clouds might not be as far off as some suspect.
I haven't sprung to read the article, but once I can find a free copy I probably will. Until then Watts Up With That, the Hockey Sctick and a few other more skeptic types of sites have reports that I really can't confirm. Anywho, the interesting thing to me is that the models might be repairable. The question though is if the modelers will actually read and incorporate these newer developments or just continue with their old song and dance. Since that might involve eating some crow, I wouldn't hold my breath.