Climate science does itself a disservice by being so imprecise and not controlling for unrelated variables. And yet your comment is flagged, and probably will be removed.
A friend's whole PhD as a physicist was to improve modeling cloud formation at a molecular level to be used in a model which was then part of an ensemble, that took him 5-6 years to manage to improve one particular parameter of a model with thousands of parameters.
It's an attempt to simulate all the energy transfers happening from solar activity on Earth's atmosphere, through all chemistry and biological reactions that contribute to results.