Hello Dennis,

  I found the cbc bug - and you probably have as well by now.

The may01-02 hindcast file is 49 records long instead
of 48 hours.  If you remove the last hour of data from
that file, the M2 signal falls out completely when you detide
(using the codes we wrote back when) and you recover
the desired clean wind-forced-only elevation.

I am not mailing this to Rick at present.  His file headers
were correct in that he had a "49" for the number of hour-
blocks in the file.  We need to revise our detide code to
check for this information. 

Dang, dang, dang.

Cheers,
Cisco.