Looking at the relationship between speed, flow and power dissipation, I come to a different conclusion.  The pump output curve goes up relatively linearly with speed  and bends over toward flat as the back pressure builds.  The power dissipation required (with a prop) goes up roughly as the cube of the speed (parabolically), so it goes up slowly at first and then heads up pretty steep as you get past 3-4000 rpm.

        Because the heat transfer behavior of the radiator and the engine block are NOT simple functions of water flow, you really gain no insight by doing simple flow calculations. You have to do the experiment, or do the computer model of the whole system. In reality, you would end up doing both.

        The heat transfer formulas are very complicated and VERY non-linear.

        If you think about the extremes, the impossibility of an accurate, but simple, calculation of heat transfer as a function of water flow becomes apparent. For example, if the airplane is in the chocks, and there is little airflow through the radiator, changing the water flow (within reasonable bounds) is not going to significantly alter the amount of engine cooling. Conversely, at any given airspeed, there is a water flow that, above which, there will be no additional cooling provided. In both of these cases, the radiator is close to the same temperature everywhere and the limitation is the heat transfer to the air, not the heat transfer from the water to the radiator wall (or from the engine to the water.)