Interactions of the shape of a particle size distribution and physical processes using a three moment snow growth model in two snowfall cases

The formation of snowfall is a consequence of interactions between physical processes and the size distribution of snow particles. In this study, 1-dimensional steady-state model in combination with weather radar observations was used to retrieve the vertical evolution of a snow particle size distribution. The snow growth model uses the zeroth, the first and the second mass moments of a particle size distribution for calculating the evolution of distribution parameters. The model includes physical snow growth processes, such as ice nucleation, water vapor deposition and aggregation. Other processes, such riming, breakup and ice multiplication were not considered in this study. Given the model setup, snowfall cases where either supercooled liquid water is absent or supercooled liquid drops are small enough that they don't collide with ice particles can be physically represented by the model developed in this study. To analyze the model performance and its ability to retrieve snow microphysical properties, two distinct snowfall cases, that took place on 26th and 30th December 2010, were selected. The December 30th case is a typical snowfall event and the model was able to reproduce particle size distributions and radar variables in December 30th case that agreed well with observations. Particular radar signatures that have recently attracted a lot of attention in the scientific literature took place during the December 26th case. It is believed that they are associated with enhanced vapor deposition and aggregation growth of ice particles. The case of 26th December could not be accurately modeled probably because of invalid steady state assumption or missing microphysical processes. The number concentration of small particles at the ground level is underestimated in model results. The model also cannot produce approximately exponential size distribution as was observed by ground instruments. The underestimated number of small particles refers that breakup of particles or ice multiplication could have a significant role in the evolution of size distribution. However, based on the model runs with different types of snow particles, the model produced high intensity snowfall when high number concentration of plate crystals existed, which could cause the above-mentioned signatures. So our study does corroborate results presented in literature, but may indicate that there is a missing snow growth process that is currently not considered. As a part of this study the currently available 2-moment model was extended to a 3-moment scheme. These two schemes were extensively tested and analyzed. Changes in size distribution shape due to aggregation and deposition is captured by the 3-moment scheme. In the 2-moment scheme shape of a particle distribution is constant. It was found that ice water content (IWC) and snowfall rate grow faster in the 3-moment model than in the 2-moment model. Fast increase in IWC due to ice particle growth by water vapor deposition leads to narrowing of a size distribution. In upper parts of clouds, the 3-moment scheme produces questionably narrow size distributions, which needs to be studied in the future.