Ice nucleation plays a critical role in forming ice in clouds and converting liquid to ice in the mixed-phase clouds, which, in turn, has important climate consequence. Many past heterogeneous ice nucleation parameterizations were developed based on field measurements with artifacts of shattering. Recent lab and field measurements led to a few new ice nucleation parameterizations, which need to be implemented to models and evaluated by observations. In this study, we implemented three very recent ice nucleation parameterizations (i.e., DeMott et al., 2014; Phillips et al., 2013; and Niemand et al. 2012) into WRF-CAM5 for mixed-phase clouds to compare with the Meyers et al. 1992 and Diehl and Wurzler 2004 schemes, and evaluated with an observed storm case during DC3 field campaign. The ice nucleation schemes were connected with dust to examine dust effects as ice nuclei (IN). We find that the three recent parameterizations predict converged results in ice number concentrations (Ni), which are lower than Meyers parameterization. The homogeneous aerosol and droplet freezing is significantly enhanced in the three schemes that give lower ice nucleation rates in the mixed-phase regime, leading to very different anvil cloud micro- and macro-physical properties due to changes on humidity and droplet nucleation. Increasing dust leads to ice-rich clouds and diminishes supercooled water, impacting cloud radiative forcing significantly.