It is well-known among climatologists that large swaths of the physics in GCMs are not well understood.31 Where the uncertainty is significant GCMs have “parameters,” which are best judgments for how certain climate processes work. General Circulation Models have dozens of parameters and possibly a million variables,32 and all of them have some sort of error or uncertainty.
A proper assessment of their physical reliability would include propagating all the parameter uncertainties through the GCMs, and then reporting the total uncertainty.33 I have looked in vain for such a study. No one seems to ever have directly assessed the total physical reliability of a GCM by propagating the parameter uncertainties through it. In the usual physical sciences, an analysis like this is required practice. But not in GCM science, apparently, and so the same people who express alarm about future warming disregard their own profound ignorance.