Dissipative particle dynamics (DPD) is a novel particle method for mesoscale modeling of complex fluids. DPD particles are often thought to represent packets of real atoms, and the physical scale probed in DPD models are determined by the mapping of DPD variables to the corresponding physical quantities. However, the nonuniqueness of such mapping has led to difficulties in setting up simulations to mimic real systems and in interpreting results. For modeling transport phenomena, where thermal fluctuations are important (e.g., fluctuating hydrodynamics), an area particularly suited for DPD method, we propose that DPD fluid particles should be viewed as only (1) to provide a medium in which the momentum and energy are transferred according to the hydrodynamic laws and (2) to provide objects immersed in the DPD fluids the proper random ``kicks'' such that these objects exhibit correct fluctuation behaviors at the macroscopic scale. We show that, in such a case, the choice of system temperature and mapping of DPD scales to physical scales are uniquely determined by the level of coarse graining and properties of DPD fluids. We also verified that DPD simulation can reproduce the macroscopic effects of thermal fluctuation in particulate suspension by showing that the Brownian diffusion of solid particles can be computed in DPD simulations with good accuracy.