So how much of the 2007 Arctic ice fluctuation was caused by CO2, and how much by natural variations in clouds, wind, etc?

RealClimate: The A-train

Among the results from the A-Train, according to L’Ecuyer and Jiang, is that a 16% decrease in the summer-time cloudcover over the Arctic played a significant role the well-known September 2007 Arctic sea-ice minimum. This comes in addition to the anomalous winds, and was news to me. But it makes sense: less cloudiness has resulted in greater absorption of short-wave radiation (sunlight) at the surface, providing increased heat for ice melting.