Not to freak y'all out, but here are three spaghetti ensemble model* current looks that include what is probably the worst case eventuality.
1. The HMNI model (an extrapolation of the HMON, my favorites) takes Irma into the Gulf of Mexico as a Cat 4+. This wouldn't be good for S. FL, the Keys, or Cuba.
2. The HMNI intensity model has Irma ramping up to a Category 5+ once in the Gulf, with sustained winds near 160Kts, 184Mph (Cat 6?...).
3. The GDF/Canadian ensemble tracks that take Irma through the South and past KY, TN, IN, OH, MI, PA, and into Canada and possibly the NE. There was another cyclone that hit the Gulf Coast and all the way to Montreal and NE some years back. Of course billions of tons of precipitation will go along.
* All the models have a disclaimer that the most current NHC forecasts supersede all, as the official position/view.
(PS: Wish the forum commenting warned of 'too large file sizes' before the Preview, I lost my former better written comment (and time) when I flubbed the error from using larger png files instead of small jpgs, grrrr. I usually take a 'copy', but didn't; that's always when there's a glitch.)
Three spaghetti ensemble images:
1. The current spaghetti ensemble, showing the HMNI entering the GOM:
[ You are not allowed to view attachments ]
2. The Intensity spaghetti ensemble, with the HMNI going to near 160Kts after Irma enters the GOM:
[ You are not allowed to view attachments ]
3. The spaghetti ensemble showing tracks through the S and mid-section, Great Lakes, Canada, NE:
[ You are not allowed to view attachments ]