00z NAM continues trend of a significant FL Keys Threat

[Vortex]

The Nam for the last several runs has insisted that Dennis would move wnw/nw beginning this afternoon. Thus far it's performing very well. It's only out 24 hours but the method behind the madness and the current synoptic environment really does make sense...We'll see how the remaining run turns out.

http://www.nco.ncep.noaa.gov/pmb/nwprod ... loop.shtml

[ericinmia]

The Nam for the last several runs has insisted that Dennis would move wnw/nw beginning this afternoon. Thus far it's performing very well. It's only out 24 hours but the method behind the madness and the current synoptic environment really does make sense...We'll see how the remaining run turns out.

http://www.nco.ncep.noaa.gov/pmb/nwprod ... loop.shtml

Some variation runs of the gfs pull a similar track...

http://moe.met.fsu.edu/cyclonephase/com ... 12/M3.html
-Eric

[Ixolib]

That's a pretty abrupt northern then western turn!!

[jax]

It also has the presure at 1008 - 1012 and implies it
only a VERY broad low... did you see the smaller low
split off and go south of Jamaica? What was that about?
I may be wrong... but I think you'er looking at the wrong
model.

[PurdueWx80]

I'm sorry, but we can't put any merit in a model that has too small a domain to even cover the initial location of the storm. This is why it handles Dennis so poorly. It will get better as we go through tomorrow (becasue the storm will be within the bounds of where the model is actually run at initialization), but for now all should be ignored. The NAM did such a wonderful job with Cindy because that storm actually formed will within the NAM's domain. The key point here is that the NAM is a mesoscale regional model that covers much of North America, whereas the GFS, ECMWF, UKMET and GEM are all global models, with domains that cover most/all of the planet.

[reeef]

one of those takes it to texas...

[jkt21787]

DO NOT take the NAM seriously at all, Purdue laid out all the problems with it very well.

[btangy]

The key point here is that the NAM is a mesoscale regional model that covers much of North America

The dashed outline is the ETA/NAM domain. Hence, it isn't used for TC forecasting very often because the model is designed to predict weather on the North American continents. The limited domain lets the model be run at higher resolution, but as you probably noticed, a shorter duration (84 hours currently) before the boundary conditions catch up and forecast skill is lost.

The closer you are to the boundaries of the domain, the less useful the model. I would only use the NAM in the very near term.