For some unknown reason this blogsite, Typepad, treated the following comment as SPAM when it was posted to the Scientists blog below. It isn't SPAM, it came from my wonderful friend Hans PD who is a scientist with a great deal of experience in environmental and atmospheric issues.
"Thanks for calling my attention to the UN letter.
"This website will also be of interest to you. It’s a long-winded way of saying what I’ve been saying for a few years: GCM’s (global climate models) are crap. When I worked with them, and what I’ve seen, and what this article states, is that they don’t predict the present, let alone the future. The funny thing for me was watching their users stand up and talk about their results as being really good, when at the same time they were showing the data/model results comparisons and they were poor.
"For example, on local pollution modeling (different than GCM) most models are happy to get a diurnal profile, even better if the highs are during the day, but almost nobody tries to get the actual pollutant concentrations correct. That’s what I did. And even that is a low standard compared to what one would call “agreement” between a theory and data in chemistry/physics/math.
Yet people would stand up at conferences showing model results w/o a proper diurnal profile (e.g. high pollutant concentrations at night, low during the day, or something else totally wrong), and say the model was working well become it showed some sort of variation. It was sort of funny if there weren’t so many blow-hards."
Let me support what Hans says. I worked for years with equations and modeling in economics. No matter how good the data, and in economics we had a lot more relevant and better data than the climatologists have, our equations could not perform better than a "straight-line" assumption for data that was prior to or after the data used in the equations. In other words, assuming that "things would continue as they were" was statistically more accurate than our complex, computer modeled equations.