I wouldn't recommend using the GFS or any other long range model as your main guidance at this point since you're within less than 24 hours of the event effecting your area. I would focus a little more on the short range models like the NAM, RPM, SERF, RAP, etc. since they tend to pick up on some of the mesoscale features that the GFS, ECMWF, CMC, etc might not see.