Synergie eines effizienten polarimetrischen Radarvorwärtsoperators und eines hochentwickelten Klassifizierungsalgorithmus zur verbesserten Repräsentierung von Hydrometeoren im ICON-Modell (Operation Hydrometeors Part II)

点击翻译按钮获取中文摘要

In Phase 1 ‘Operation Hydrometeors’ provided an efficient polarimetric forward operator as a fundamental prerequisite to the Special Priority Programme required for the fusion of radar polarimetry and atmospheric modelling. The non-polarimetric Efficient Modular VOlume RADar OPerato (EMVORADO) has been extended to polarimetry assuming oblate spheroids at precipitation radar wavelengths in online implementations in the NWP-models COSMO and ICON and a stand-alone framework. We also developed a sophisticated and more robust hydrometeor classification (HMC) algorithm based on clustering with the additional capability to estimate hydrometeor partitioning ratios from both polarimetric radar observations and synthetic variables based on ICON simulations applying EMVORADO. With the tools developed, we evaluated polarimetric signatures and hydrometeors in stratiform rain employing a dual strategy, based on advanced radar-based hydrometeor typing and quantification and a direct comparison of multivariate observed and synthetic polarimetric signal distributions. Some improvements of the two-moment cloud microphysics parameterization in ICON resulted from this already.Phase 2 further refines the developed tools to broaden their applicability. We improve the melting scheme in EMVORADO, potentially also interface a scattering data base, and refine our new HMC to enable hydrometeor clusters to extend across the 0°C level. With these developments we evaluate and improve with our dual strategy the representation of three more key polarimetric signatures pointing to different model deficiencies: (1) Columns of enhanced differential reflectivity ZDR will guide improvements in the freezing process of raindrops, (2) ZDR near the surface is the key sign for a better parameterization of raindrop size distributions, and (3) polarimetric variables in the dendritic growth layer show pathways to improve and extend the representation of ice microphysical processes (e.g. secondary ice production) in ICON. Finally, we will exploit polarimetry-derived information for indirect data assimilation in ICON. Based on the developments of the project, we assimilate ZDR-column objects and/ or the partitioning ratios derived from our new HMC, most likely in the form of categorical ranking information.