Evaluating Ice Phase Microphysics in the Simulation of a Snowstorm Over Northern China

The complexity of ice particles in the atmosphere makes it difficult to model microphysical growth processes accurately. In this study, we simulated a snowfall case over Northern China Plain using two different microphysics schemes, that is, Thompson and Morrison schemes, in the Advanced Research WRF (Weather Research and Forecasting) model. Both schemes are able to reproduce the event, albeit with a slightly weaker precipitation compared with the surface observation. However, the radar reflectivity factor in Morrison simulation is higher than the radar observation to similar to 10 dBZ. Further analysis reveals that such stronger radar reflectivity in the Morrison simulation might be caused by larger collection efficiency, which would lead to more active self-aggregation process in prediction of snow number concentration and then larger snow particle size. Sensitivity tests show that using an alternative formula of collection efficiency produces smaller radar reflectivity that is in better agreement with observations. This study highlights the accurate representation of self-aggregation process and underscores the needs of further improvement of ice microphysics schemes for the better snowfall simulations. Understanding how snow forms in the atmosphere is complicated. We did a case study using two different microphysics schemes in numerical weather model to simulate a snowfall event over Northern China. Both schemes did a pretty good job in simulating the snowfall, but one of them, the Morrison scheme, showed much larger radar signal compared to what was really happening. We do further analysis to find out why this was the case. It turned out that the Morrison scheme simulated the snowflakes stick together more than they actually do in nature. This made the size of snowflakes in the simulation much larger. We ran more tests by changing the way the Morrison scheme making snowflakes stick together. When we did this, the simulation results showed more accurate radar reflectivity factor. So, what did we learn? How the snowflakes stick together is a big deal. If we get it wrong, the models will give us the unreasonable estimation of radar signal. This study helps remind scientists that they need to be very careful when they use these microphysics schemes. Radar reflectivity factor from Morrison simulation exceeds 10 dBZ compared with observations, but Thompson scheme performs better The overestimation of radar reflectivity factor in Morrison simulation could be attributed to larger snow particles Larger snow particles in Morrison simulation are caused by active self-aggregation process of snow and higher collection efficiency