Plot of the raw data.

     Process SWELL computes the average water bottom time of a group of 
traces, then shifts the middle trace of the group by the difference of
the middle trace water bottom time and the average water bottom time.
     Process WBT has several automatic water bottom time pickers.   
1)  The simplist picker is to pick the largest amplitude in each trace.
Unfortunately, the largest amplitude jumps around in this dataset.
2)  The short-over-long window (solrat) method also fails for the same
reason.  (Normally the long window is set to a window well after the
water bottom, but in this case the amplitudes frequently are higher later 
(basalt basement?).
    Filtering before picking with PEAK or SOLRAT didn't help.

    The THRES method picks the first trace amplitude that exceeds the
THRES value.  Using process PROUT with parameter    sets .06 .07
thres .3E-08 was chosen but had to be lowered since many traces didn't
have a water bottom amplitude exceeding it.  thres .1E-08 was chosen, and
track .002 was used to keep the pick from wandering on "bad" traces.
    Notice that the swell period is around 50 traces.

Plot with WBT track .002 thres .1E-08 and SWELL n 50 
Plot as above but with FILTER and different grayscale.

Scripts - all the scripts used.