Just in case this helps someone else out, the issues I have been having exporting the NHD flowline data from postgre is actually because of a memory leak in the pgsql2shp binary. I have looked at the source code, ran the process with my debugger attached, etc... and that thing throws a heap corruption for every single row it exports. I have given this info to the postgre and postgis folks and re-compiled postgis from source.
Seriously though, I'm at the point where if anyone has the already processed flowline shp file I'd be willing to find a way to compensate anyone who could upload it to my ftp server. I could give you some free ftp storage or something. I don't know why you all are able to export the data without all of these problems, but the memory leak is there, so either you are crunching less data so you don't run out, or you have more RAM than me or something. I still can't believe how much code I've had to look at so far for this map, lol. The only thing my map needs is the completed flowline data.

Anyway, however it works out I'll probably take a break for a week, and start on Oregon sometime the first week of February.