In preparation of its texts, IPCC utilizes General Circulation Models (GCMs), analyzed using computers. Such models, and the way in which IPCC utilizes them, exhibit — among others — the following errors.
There is no mathematical apparatus (i.e., proven theorems) behind these models and “ensembles”. This fact makes the whole modeling exercise useless. This fact also makes rigorous evaluation or criticism of these models very difficult or even impossible to do. The following notes apply to the GCM models (including their more complex variations, such as AOGCM).
GCMs are in violation of the most basic law of the information theory. GCMs purport to create information from nothing. Meteorologists, using accurate data and valid weather models, integrate them on a high resolution space-time grid, and utilizing statistical regression, can forecast the weather up to two weeks in advance. Forecasts for longer periods cannot be made because the loss of information is exponential with time. This is a property of valid weather models, which must comply with more basic laws of fluid dynamics. Running a valid weather model for years (assuming that were possible) would result in an end state having near-zero information (in the sense of the information theory). Averaging even a very large number of such results would still yield near-zero information. Any “method” alleged to recreate the lost information is either a fallacy or a fraud, comparable to a perpetual motion machine of the second kind. (In fact, the “climate modelers” use semi-random data and dubious weather models, integrate them on a low resolution grid, and have no benefit of statistical regression; although that hardly matters given all other errors.)
IPCC utilizes “ensemble of models”, apparently relying on the central limit theorem of the probability theory. One of the results of this theorem is that average of a large number of independent measurements or estimates converges to the true value, subject to some conditions. But the models are not independent, so their average does not converge to the true value, if it exists at all.
These errors in application of mathematics and computer sciences are in addition to the IPCC errors and frauds in physics, biology, and other sciences.
There is a very interesting result published by scientists from the Prague University (founded 1348), rigorously quantifying dishonesty of literature on the climate sensitivity. This dishonesty is in addition to the modeling errors.
Revised on 03/01/2017.