Multifluid Simulation

The short answer is no.  Flow Simulation is limited in the differences in densities of the "fluids" in a simulation.  It does not do mixed phase (gas and fluid interaction).  For some problems you can get an approximate solution by using fluids of differing densities but with the rotating impeller it adds a degree of complexity that will likely overpower the effects of this crude workaround.

This is unfortunately well beyond the capabilities of Flow Simulation. Multiple fluids can be used, but they must be the same phase to touch one another or separated by a solid medium. Particle studies can predict what small amounts of gas or solid debris will do inside your flow, but their use is limited; they will never affect the flow of the liquid itself, for example. Also you mention the turbine is only partially submerged, so you would end up throwing water and cavitating water constantly; this is an extremely complex multiphysics problem and even the most advanced solvers would have difficulty solving it. Consider that for a 2D, one cell thick cavitation problem using water and no moving parts, it takes about an hour for an analysis to predict what happens. A rotating fan blade moving just air takes 1.5 hours. Now you have multiple complex shapes moving water and air, free-surface liquid, cavitation, and high speed flow at different densities all at once. Not a pretty picture from an FEA standpoint. The purpose of FEA, currently, is not to exactly replicate real life but to analysis specific scenarios to gain insight one variable at a time.