Okay...let me take a crack at this...the only models y'all need to pay attention to are the GFS, ECMWF, GFDL and the HWRO; three of them can be found at the Models and Forecast Branch over at NCO NCEP's website and I doubt you're going to find another website with the resolution they provide. I use San Diego State University's Weather Lab which overlays the GFS and ECMWF on one chart.

The NAM is also found on the NCO NCEP website and has been around as far back as my active duty in the Air Force but under a different name and is also reliable; however like all models each because of what Ed just wrote about use different formulas and continue to be tweaked to this day.

This year, the NAM has had a bias for depicting systems in the low latitudes as being stronger than other models depict them to be. The GFS and ECMWF have been neck and neck this year and have performed exceptional over the Atlantic despite no upper air data.

As a matter of preference I analyze the 00Z and 12Z packages because those charts have been updated to include new upper air observations that launch at those times across the world.

I can't speak for Ed but you may have noticed we don't get caught up in every little detail of each of these charts; what we're looking for is continuity. We locate where everything and like many of the discussions we're looking for changes and what chart is performing the best. We look at everything from satellite imagery, streamlines, shear analysis and depending on where we make our comments depends on what forum we're in.

Each of you love weather or y'all wouldn't be here. It's my job to try and articulate to the best of my ability what it is y'all are looking at, talking about and provide assistance. We certainly don't want to talk over your heads.

I do stress this caution....and Ed's dead on about this....the model outputs are for the most part very good out to 120 hours. There are some scenarios we can go out further and others as few as 48 hours and sometimes even less.

Tonight, Hurricane Danielle is pretty straightforward however at the end of the period the models do not agree on where everything should be and I cannot stress enough how important these little differences are; particularly for Bermuda and the NE United States. I can tell you we think Danielle will re-curve and we were pretty confident about that two days agao. We're not quite there tonight because the data we all look at isn't clear cut nor in agreement with each other.

Tropical Storm Earl while it appears it's making a carbon copy trace of Danielle will likely stay on a more southerly track and we don't know where recurvature is likely to be at this time. We think it will and get hung out to dry south of Bermuda and well offshore NC/SC.

I will do my best to answer all of your questions, but for the next few days; maybe at least five days, there's not much we're going to know on Earl until the storm is where Danielle is now.

Finally, we don't know if a model depicts a low that be default makes it a tropical cyclone. What our eyes see does not imply it is what we initially think or want it to be. I reference the NAM and the Gulf of Mexico these past couple of days. At first glance, you'd swear there's something to it, but without everything else we simply can't jump on the band wagon and say; that's a tropical cyclone and hit the panic button.

There will be data and models each of you are going to look at that we may dismiss and I promise y'all it isn't deliberate.

Tonight...both the GFS and ECMWF are doing exceptionally well with Danielle and we have reason to believe they are the models of choice for Earl in the days ahead. These models are the big boys...their global models. The GFDL as Ed so eloquently stated is on a smaller scale and really zeroes in on that storm's environmental envelope.

We use everything in concert with each other. We also rely on y'all to help us zero in on what each of you are looking at; thus saving us time to analyze and get back with each and all of you.

Have a good day/evening...

Bill