Constraints on spatial distributions of negative energy