Hello,
This is explained page 181 and 182 of my book "hydrodynamics of free surface flows" published by Wiley. In the classical form the upwinding does not depend on the CFL number, in the modified form it is proportional to the CFL number (it tends to a centred scheme when there is no velocity). I cannot give a pdf of my book because it is sold by Wiley, but it seems that a number of websites have it and others allow to browse it with only a small number of pages missing.
As for the reference of Viollet, Munk and Anderson, I cannot find it, it seems to be lost.
With best regards,
Jean-Michel Hervouet