A new paper published in Geophysical Research Letters demonstrates the pathetic state of climate modelling of clouds. According to the paper, the only way for the modellers to reproduce cloud effects on observed temperatures was to use microphysical properties of clouds that were the most divergent from satellite observations. In other words, to “tune” the model to reproduce one observation [temperature] results in detuning and non-reproduction of another observation [cloud microphysical properties]. The models thus remain in a state of infancy regarding clouds, one of the most important parameters required for climate projections. A mere 1 to 2% cloud modelling error can alone account for global warming or cooling, and these computer modelling games are nowhere close to achieving such a level of accuracy.
The models also spectacularly fail on many other vital components of climate.