The west coast of Florida was at the center of the predictions two daya ago (about when it hit Cuba), which is OK for most purposes. Don't want to evacuate EVERYONE for hundreds of miles, every time a blow comes up.
Hurricane danger is pretty well known in decade timescales (to support insurance, city panning and backup generators/water stocks), and on days timescales (computers and satellite observations, among other things). No particular reason we need monthly or weekly timescales covered.
When Mt. St. Helens blew, there were predictions on the 'within a century or two' timescale from geologic and historic observations, and predictions on the 'within a few months' from seismic observations of magma movement, but no one knew on the days scale. When the big blow came up, there was about an hours' time when you could rush away from the eruption and hope to beat the ash cloud, but no real pre-warning.
So, you use the months-time-scale information, and evacuate the 'red zone', then warn and post alarms in the surrounding area, and put the Guard on alert. Dixy Lee Ray kinda had it right.
Hurricanes, on the other hand, you get lots of useful timescales covered by the forecaster techniques. No problem with the computers at all.