Can anybody chime in on the relative reliability of the various models?
The GFDL course has me kinda uncomfortable, in my unscathed-by-Katrina suburban New Orleans home. I think it's time to replace the boards for my windows tomorrow (beat the rush).
The models are in very good agreement out to three days or so, given the strength of the high to the north of Dean. If anything, the intensity forecast is conservative, given the tremendous deep warm water in the western Caribbean now.
However, I wouldn't put much credence in any model past 72 hours. After all, Rita 96 hours out was forecast to hit Matagorda Bay. It hit the TX-LA border instead.