A Comparative Analysis of a 5-Year Statistic of Polarimetric X-Band Radar Data with ICON Simulations

This study presents a 10-year climatology of quasi-vertical profiles (QVPs) of polarimetric variables and ice-microphysical retrievals, such as ice water content (IWC), total number concentration (Nt) and mean volume diameter (Dm) in stratiform rain, monitored with the Bonn X-band radar (BoXPol) located in western Germany. Particular attention is paid to the layers of dendritic growth (DGL) and melting (ML). Stratiform sequences are identified applying the information entropy in combination with a ML detection algorithm to QVPs. Polarimetric variables and retrievals are compared with their synthetic/modelled counterparts from the ICON numerical weather prediction model using DWD's operational radar observation operator, the Efficient Modular VOlume scan RADar Operator (EMVORADO), recently extended to polarimetry and allowing direct comparisons in radar observation space. Analysis of the differences between the observations and the modelled/synthetic counterparts reveals both deficiencies in EMVORADO and parameterisations used in ICON. In the retrieval space, too low Nt, too high Dm and too low IWC values in the simulations in comparison with their polarimetry-derived counterparts are an indication of insufficiently represented processes in the model. In the observational space, pronounced differences in the polarimetric signatures of the ML and DGL confirmed the conclusions from the retrieval space. For example, a more pronounced gradient of ZH above the ML is an indication of the more distinct aggregation and riming processes that are present in the simulations. Significantly higher maximum values of ZH and ZDR in the vertically extended melting region, and persistently high ZDR towards the ground, are further indications of intense riming processes and melting graupel.