Yeah, the problem with the air-water ICs is all in the steady-state boost conditions...with a 2nd rad, if the water is cool, you'll get good air->water heat transfer but not a good water->air x-fer. If the water is hot, air->water is bad and water->air is good. In heat x-fer you always want the temperature differential, and the water gets in the way, so to speak, of getting the heat eventually to ambient air. Which is why the fast sy-ty guys always switch their air-water IC units out for a big air-air.
Unless if this is a track only car, in which case the lower than ambient cooling possible with ice water would be great, it's not very practical.