Many years ago I had the opportunity of doing a small amount of
experimental work
in fluid dynamics, in the laboratory which Raymond Hide then ran at the
Meteorological Office.
I was working, with a colleague, on the effect of `bumps' on fluid
spin-up in a rotating
circular cylinder. We found that small changes in the boundary geometry had little effect if they were symmetric about the rotation axis,
but that non-axisymmetric changes could affect the spin-up time
quite substantially, especially at relatively high Rossby numbers (~0.25).
In particular, a bar of square cross-section, attached to the base
along a diameter, and with a height only 5 per cent of the cylinder height,
decreased the spin-up time by a factor of 7. So far as I am aware, this
substantial effect of non-axisymmetric `bumps' has never been
reported in the scientific literature, and would seem to merit further
investigation.
(Note: Extensive recent work on this problem by Richard Hewitt
at
Manchester University can now be found here.)