Most numerical models of cirrus clouds utilize a constant deposition coefficient, a quantity related to the efficiency of attachment of water vapor to an ice crystal as it grows. Another assumption made by most cirrus cloud models is that ice forms and grows in the shape of a solid sphere. Unfortunately, these assumptions are not consistent with laboratory measurements. Using a theory of ice growth consistent with lab measurements and a Lagrangian microphysical parcel model, I studied more realistic ice cloud modeling though the incorporation of a varying deposition coefficient and more realistic ice crystal shapes. Throughout my research, I found that varying the deposition coefficients of ice crystals to simulate rosettes, a more realistic ice crystal shape, produced growth different to that of the previously assumed spherical crystals. In fact, rosettes were found to grow closer to column crystals than plate crystals or spherical crystals. The rosette growth is faster, depleting the supersaturation at a greater rate, than spherical growth. Consequently, homogeneous freezing is shut off sooner and at a lower altitude. This result may explain the lower ice concentrations found observationally in cirrus clouds compared to those found in models.