======================================================================
                    R E A D M E 
                    doc: Thu Dec 30 01:20:33 2010
                    dlm: Thu Dec 30 01:24:55 2010
                    (c) 2010 A.M. Thurnherr
                    uE-Info: 34 71 NIL 0 0 72 3 2 4 NIL ofnI
======================================================================

This is the version that I arrived at after implementing a surface-wave
correction based on CTD acceleration. This version allows the
correction to be specified as a polynomial of arbitrary order,
separately for the down and upcast. The quality can be assessed by
comparison with bottom-tracked w profiles and, to a lesser degree, by
comparison with the "diagnostic w profiles" derived with the shear
method. Some preliminary findings based, primarily but not
exclusively, on the P403 yoyo data:

1) The surface-wave acceleration correction considerably improves the
cast-to-cast consistency of vertical velocity profiles.

2) It appears better to use a single correction derived from several
casts, rather than a cast-based correction.

3) A linear fit seems to work better than a higher-order fit.

4) Based on DIMES UK2 stations, it appears that a near-constant winch
speed is required for this to work.

5) The method appears not to work for data sets collected with large
rosettes. This is presumably due to package-induced pressure effects.

Since the best validity test I have been able to come up with is based
on bottom tracked w profiles, I decided to fold the BT code from the
shear method into this code. I save the current version, just in case.