zdas04
Mechanical
- Jun 25, 2002
- 10,274
Yep, that's exactly what they're saying. The hindcast concept only exists in climate science. They input all the parameters (including that there was a huricane) and the model tells them that there should have been a huricane.
It gets even flakier when they put in the whole storm track and tell the model to track the storm--that exercise has maybe a 30% success rate with model following the track.
Weather models can't reliably predict what is going to happen in the next 6 hours (I live in the Rockies and the effects of the mountains and valleys are just too complex for the models they use). A weather model looks at an area about 1/4 the size of the continental U.S. and can't predict this afternoon. I'll leave you to estimate the "accuracy" of the models that try to predict global performance for the next 50 years.
David
It gets even flakier when they put in the whole storm track and tell the model to track the storm--that exercise has maybe a 30% success rate with model following the track.
Weather models can't reliably predict what is going to happen in the next 6 hours (I live in the Rockies and the effects of the mountains and valleys are just too complex for the models they use). A weather model looks at an area about 1/4 the size of the continental U.S. and can't predict this afternoon. I'll leave you to estimate the "accuracy" of the models that try to predict global performance for the next 50 years.
David