The central importance of feedbacks to the global warming issue can not be overstated. How the radiant flows of energy in and out of the Earth system change with temperature is THE most critical piece of knowledge we need in order to predict whether manmade global warming will be benign — or catastrophic. It is obvious that good estimates of feedbacks are needed from our observations of natural climate variability. This would provide the most important test of climate models: do the models exhibit feedbacks consistent with those we observe in nature?
I am now convinced that the main difficulty in diagnosing feedbacks from observations of natural climate variability has been related to the issue of causation when observing cloud behavior. The best way to introduce what I mean by this is with the following question:
When low cloud cover is observed to decrease with warming, is the cloud change the result of the warming, or is the warming the result of the cloud change?