While it can be done, I don't think it is hardly the easiest way to do it. I don't think it would be possible with commercial though either. Simulating fluid dynamits is hard. It is even hard for software which is set up to calculate a short segment of time over a long period of time (determine what will happen when a landing craft hits the water at 50 mph...that 30 second or so time period takes well over a days worth of CPU time, and is still inaccurate).

You should be able to "fake it" some and get satisfactory results using the method you are persuing right now. This post touches on some ways to use the physics engine with water though.

http://www.conitecserver.com/ubbthreads/showflat.php?Cat=&Number=288647&page=1&view=collapsed&sb=5&o=&fpart=1