E_COEFFICIENT was a concept I introduced in FDS even before it was FDS. We needed something to model fire suppression. At the time, there was some research in developing empirical models of suppression, all of which involved some sort of exponential decay assumption. There is some physical validity to this assumption, but the concept of E_COEFFICIENT (and all other suppression models) never really gained traction.
If you want to model suppression of a real fire burning real stuff, I strongly suggest that you get some real stuff and burn it under a calorimeter with different amounts of water applied. I cannot imagine how you could determine an appropriate E_COEFFICIENT any other way. It is a value that would be very difficult to tabulate in a handbook.