danielw -- indeed it is.

To illustrate this, I used one of the utilities we have here called GEMPAK. Those of you in meteorology are probably familiar with it; for those of you who aren't, simply put, it allows you to display a whole bunch of meteorological data. It's akin to a programming language to display data.

Since it is a rather large file, I have uploaded it to my webserver so as not to clog this webserver down. It is in Adobe PDF format and is about 900kb in size. The file contains a plot of all 12z radio observations from across the eastern US & Bermuda along with three model output runs for comparison -- GFS in lime green, NOGAPS in red, and Eta in a barely readable yellow (my apologies).

You can get the file here:
http://www.northflams.org/misc/500height.pdf

Contours are denoted every 20m, i.e. 5900m, 5920m, and so on. The number in the upper right of each station obervation gives that station's height. Of particular note are how the models handled the height in Bermuda and the heights along the east coast in Charleston and Jacksonville. Only the GFS came close; the other models kept the 5900 and 5920 lines well out to sea. I could do the same for other times and show the same thing -- and this is all without the aid of the radio observations from the G-IV surveillance plane.

Just an interesting piece of info to leave you all with for the evening.