Nice post. Interesting. I'll look into it. Here are a few notes after a quick (NOT in depth) review of the IPCC chapter 8.pdf you referenced...
1. “Some models used for projections of tropical cyclone changes can simulate successfully the observed frequency and distribution of tropical cyclones.”
Really? Then why have they missed the boat in the number and severity of Atlantic hurricanes for the last several years? Or are cyclones and hurricanes really that different (from a climate model perspective)?
2. “The possibility that metrics based on observations might be used to constrain model projections for climate change has been explored for the first time... a proven set of model metrics that might be used to narrow the range of plausible climate projections has yet to be developed.”
Am I reading this wrong - aren't they admitting they DO NOT use Real World observations to refine the models? If they haven't been doing this (i.e., “explored for the first time”), just what criteria have they been used to validate the models? And just who defines what is “plausible”?
3. “Eighteen modelling groups performed a set of coordinated, standard experiments, and the resulting model output, analysed by hundreds of researchers worldwide, forms the basis for much of the current IPCC assessment of model results.”
What's wrong with just comparing the model predictions against observed Real World data? And as for noting it was reviewed by "hundreds", how many folks does it take to disprove a theory? (answer: ONE) Were these folks TRYING to find problems, or looking justify the conclusion?
4. “The response to global warming of deep convective clouds is also a substantial source of uncertainty in projections... and it is not yet possible to determine which estimates of the climate change cloud feedbacks are the most reliable.”
Doesn't this say they DO NOT fully understand - and the models do not agree on - what they admit is a key component of accurate climate change prediction?
5. [section 18.104.22.168]
This appears to be a caveat claiming that if there are differences between the models and observations, the differences are 'insignificant' and can be ignored. It also explicitly avoids doing one of the tests I suggested, e.g., predict the 20th century based on the preceding 500 years of raw data. Is that such a bad idea? Why?
* * * * *
It's a long report, but I'll continue looking in depth - I've looked at others from IPCC and was not impressed with the science, but was stunned at the numerous justifications for why I should accept the conclusions anyway.
The graph you supplied is impressive, but I don't see a disclosure on where the models got their raw data, e.g., if you're predicting 1965 based on 1964 (or 1920!) measurements, that's one thing. But to 'simulate' 1965 based on 'estimates' is something completely different. To properly evaluate the process that defines where you finished, you need to know precisely where you started.
*sigh* I cannot concede your point (yet), but I am open to the idea. Caveat: I must admit that - given the nature and scope of the manipulated data by CRU - I have doubts about the accuracy of models used by the IPCC report, simply because the integrity of the DATA is now in question. And I'm less concerned about the content of the CRU emails than I am about the accuracy of the raw data that has been made available to researchers. My simplistic review of source code segments *clearly* shows the input to the underlying model code was 'tweaked' in order to generate a specific result. Not good.
BTW - According to the IPCC, carbon dioxide causes 1.1 degress C of warming if it DOUBLES. The real key to defining the cause of climate change is feedback based on water (vapor) - and as I pointed out in #4 above, they admit they don't have a good handle on it.
More to come.