Wednesday, October 05, 2016

Matthew's Morning Models















#11,796


The National Hurricane Center's task of forecasting the track of Hurricane Matthew hasn't gotten any easier with this morning's computer model runs, as they are even less in accord than they were 24 hours ago.

When a hurricane parallels the coastline, as Matthew is expected to do, then a shift of even 50 miles - one way or the other - makes a world of difference in how the effects will be felt on land. 

And while some of the model runs (including GFS and the EURO) show a landfall on the east coast of Florida, others in the ensemble keep the storm far enough off the coast to greatly reduce its impact.


Some of the models even have the storm hitting, then turning back into the Atlantic, only to loop back to the coast.  While unlikely, hurricanes have done stranger things in the past . . . .

As George E. P. Box, Professor Emeritus of Statistics at the University of Wisconsin, famously said:

 “All models are wrong, but some models are useful.”

While not in complete agreement , the models we have are useful enough to tell everyone along Florida's east coast - from Miami to Jacksonville - that they should be making preparations for a very unwelcome visitor.


For now, the NHC is expecting the storm to either hit, or come very close to, the southeast coast of Florida and then move north.  A hurricane watch has been extended north to Jacksonville, Fl.



Hopefully we'll see some consolidation in the model runs later today, but any track parallel to the coast will be problematic for forecasters.


As always, your first stop should always be the National Hurricane Center in Miami, Florida for forecast tracks and your county Emergency Management Office for local flood maps, evacuation orders, and shelter locations.


And if you are on Twitter, you'll want to follow @FEMA, @CraigatFEMA, @NHC_Atlantic, @NHC_Pacific and @ReadyGov.