from 2 March, 2009.
Lucia writes in her post
“….there are a number of historical observations of climate that models “predict” poorly.
One of these is the average global surface temperature in non-anomaly degrees C for the entire 20th century.
My understanding is that letting people learn that models don’t predict the surface temperature accurately is considered “confusing”. One is not supposed to discuss “confusing” things as it could “confuse” people and “confused” people may come believe models based on physics may produce biased predictions of global surface temperatures in non-anomaly degrees C.”
The reason this matters is that if the average global surface temperature is biased (warm or cold), the regional temperatures are likely at least as biased. Since the impacts community needs the actual temperatures, not the anomalies from a biased average, this is a serious flaw in most of the models as Lucia, so ably summarized in her figure.
Also, in terms of predicting the outgoing longwave radiation from the surface (which is proportional to the 4th power of the absolute temperature), a -1C error, as one example from the biases displayed in Lucia’s figure, corresponds to ~1.4% error in that calculation.
Perhaps by 2011, these errors have been corrected. However, if they have not, this is another demonstration of the challenges faced by the multi-decadal IPCC type models to provide skillful predictions to the impacts community.