If you relook what I wrote more closely - I stated that if the
acceptance criteria for a GCM was to accurately predict temperature and rainfall for specific regions, and if the model was found to not be able to meet that criteria, then the model is wrong and should not be
relied upon.