Media matters has an article that says some interesting things.
http://mediamatters.org/blog/201108010025
Quote:
Spencer thinks climate models are overestimating the magnitude of the positive feedbacks, but concedes that his paper doesn't prove that to be the case. Rather, he says his study had the limited aim of responding to a previous publication by Andy Dessler of Texas A&M University which estimated cloud feedback in a way Spencer believes is faulty. "Our paper would never have been written if not for the need to answer Dessler's paper," he said.
Spencer's study compared temperature measurements from 2000-2010 to data from a NASA satellite on how much energy is leaving the atmosphere. He then compared that information to six climate models ("the three most sensitive models and the three least sensitive models") over the 1900-1999 period and found what he calls "huge discrepancies" between the models and the observations regarding the relationship between temperature changes and the energy radiated. But he added: "While this discrepancy is nominally in the direction of lower climate sensitivity of the real climate system," a variety of other factors affecting the statistics "preclude any quantitative estimate of how large the feedback difference is."
In response to Spencer's paper NCAR scientists Kevin Trenberth and John Fasullo ran their own comparison of the models and the observations and concluded that the results depend largely on which models are chosen. In other words, if you select the models in a different way from the 6 used by Spencer, the "huge discrepancies" disappear. Trenberth and Fasullo concluded that "there are some good models and some not so good," but "the net result is that the models agree within reasonable bounds with the observations." They also stated that Spencer's paper "has very basic shortcomings because no statistical significance of results, error bars or uncertainties are given either in the figures or discussed in the text."
In the paper, Spencer goes on to use what's called a simple climate model to interpret his results and concludes that you can't estimate feedback from satellite data because "natural cloud variations" driving temperature changes can be mistaken for cloud feedback. This conclusion does not speak to the validity of the climate models but rather to the validity of attempts to test the models' estimates of climate sensitivity with satellite observations. For their part, Trenberth and others take issue with Spencer's model and say temperature changes drive cloud behavior, not the other way around.
|
Still media, but it does look like the media matter authors contacted Spencer and asked him some questions.