I don't think that science needs to be grounded in anything -- it's raw observation. In fact, philosophy or metaphysics should be grounded in observation, because if it's the other way around you get biased observation. Hence, and in this context particularly, Hempel's dilemma is IMO inconsequential.