Of even more interest would be to use the model to determine what the temperature and other weather parameters were for 100 locations on earths 5-10 and 25 years ago. This would be a quick way to determine the value of the model. If it can’t “retro-predict” it’s worthless.
Of even more interest would be to use the model to determine what the temperature and other weather parameters were for 100 locations on earths 5-10 and 25 years ago. This would be a quick way to determine the value of the model. If it can’t “retro-predict” it’s worthless.True, but if I recall correctly, this has already been done a long time ago, and the models failed miserably indeed. So, they began tweaking the models’ numerous parameters until they did “predict” the past correctly. Of course, as everybody who has ever tried to use an interpolated curve to extrapolate into the future knows, this does not make the models any better. If anything, they will fail even harder.
They probably know this, or they would proudly make public predictions everywhere…