you're using an incompressible simulation for air? I've never simulated air before but that seems like quite a significant simplification, if you expect realistic results