In this paper, a novel fluid dynamics-based approach to channel modeling, which considers liquid droplets as information carriers instead of molecules in the molecular communication (MC) channel, is proposed for practical macroscale MC systems. This approach considers a two-phase flow which is generated by the interaction of droplets in liquid phase with air molecules in gas phase. Two-phase flow is combined with the signal reconstruction (SR) of the receiver (RX) to propose a channel model. The SR part of the model quantifies how the accuracy of the sensed molecular signal in its reception volume depends on the sensitivity response of the RX and the adhesion/detachment process of droplets. The proposed channel model is validated by employing experimental data.