This is a chronicle of the data cleaning efforts, which I used to fix
mistakes found in the first round of analysis.

Problem: July coverages were too small, and contained only anchoring
  vessels
Cause: wetstorage and liveaboard coverages did not have "moored"
  attribute filled in.
Solution: Filled in "moored" data set for [wl]h07XX coverages.  Saved
  edits, and exported those coverages to [wl]h07??.e00 files.  Those
  edited files were saved into exports.tar

Problem: 12-19 coverage was too big
Cause: 12-19 contained three dates: 12-19, 11-12, and 09-12
Solution: Fixed the grep statement in *modifications* to grep only for
  \1-\2-

Problem: 04-05 data set too small
Cause: Date only contained anchoring boats.  [wl]0405 coverages had
  dates of 04/04.  When these points were compared to respective
  [wl]0404 coverages, they were identical.
Solution: Drop 0405 a date.