Seems to me that yesterday the American modelers predicted routes for Dorian that came up the Florida peninsula, while the European ones predicted routes parallel to the coast to the east, over water. Today, all were predicting the latter. As I recall there was a big storm last year with a similar divergence, and after a bit, the American models echoed the European ones. So, the American modelers need to reconfigure their variables/algorithms, no?…but maybe they don’t know what they’re messing up.