The link works OK for me, so I don't know how to help you see it. Sorry
Fluid drag is indeed complex, & both viscosity & density matter.
You'll remember, Zeke, that friction factors are a complex function of
Reynolds number, Re, & that Re = rho*V*L/mu
(rho = fluid density, V = flow velocity, L = a characteristic dimension,
such as distance along the boat from the leading edge/bow or the
diameter of a pipe, & mu = fluid viscosity)
To the non-initiates, Re is the ratio of inertial to viscous forces for
the fluid against the surface, & the higher its value the more turbulent
the flow is likely to be.
Anyway, friction factor at a given velocity falls as Re increases, & Re
increases as water temperature rises, so fluid drag should fall. In the
simple case, however, we know that fluid drag does not fall very rapidly
with increase in Re, so there ain't a big gain for the relatively small
change in Re over a rower's water's likely temperature range.
Complicating things a bit more: the transition point between laminar
(low drag) & turbulent (high drag) flows over a surface moves a little
closer to the bow as Re rises, so more of the surface encounters the
higher-drag turbulent flow regime.
And then there's wave drag......
So I'd accept Kleshnev's 1.3% increase in speed between 5 & 30C as a
good empirical figure. Question is: how are you going to use this info?