The practical question, of course, is how tall of a support
can one manage on a kayak. Given the good ground conductivity,
I don't think there will be much lost efficiency in using
a quarter wave rather than half wave end-fed radiator,
but a huge difference in the effect of the required mast
swaying as the kayak rocks in the waves. A 10m mast
provides a lot of mechanical advantage compared to the
cross-sectional dimensions of a typical kayak, so it
wouldn't require a lot of sideways wind pressure to roll it.
I suspect that a more practical maximum mast height would
be 3 to 4m, which is less than a quarter wavelength. Even
at that length, an efficient base loading coil won't cause
as much loss of efficiency as it would for a ground-mounted
vertical of the same type because the earth losses are
lower. A better solution, though, would be to run the
wire to the top of the mast then angled forward down
towards the bow to make up any additional length required.
(The wire shouldn't come down lower than the midpoint
of the mast, however. If that isn't enough, some
inductive loading can be added, perhaps at the top of
the mast.)
One approach that has worked for me is to connect a base
loading coil between the antenna wire and the counterpoise(s)
after lengthening the antenna as much as is practical
up to a quarter wavelength. In this case the radial
may be shorter than a quarter wavelength also. Add
enough coil to bring the combination to resonance, then
make a coupling link from 2 or 3 turns of hookup wire
around the main coil and connect this to the coax. The
number of turns on the main coil sets the resonant
frequency, while the number of turns on the link sets
the impedance. This may not be quite as efficient as
other forms of loading, but it is fairly easy to adjust
the SWR. I've used it for a 80m shortened vertical and
it worked well.