Using Triple-Frequency Cloud Radar Observations to Evaluate Ice Microphysics in the New ICON Model

Understanding ice microphysical processes is equally challenging for modelers and observationalists because of the vast number of degrees of freedom involved. Even state-of-the-art two-moment schemes, such as implemented in the new German ICOsahedral Nonhydrostatic (ICON) model, are still facing problems to correctly reproduce basic ice microphysical processes such as depositional growth and aggregation. Retrieval algorithms based on remote sensing observations face similar challenges due the necessary assumptions in the inversion methods. In this study, we combine novel multi-frequency cloud radar observations with a versatile model framework which allows us to perform specific experiments to explore the potential sources of model-observation discrepancies. This contribution aims to exemplarily show how we can utilize the rich microphysical fingerprint in modern remote sensing observations to evaluate and improve the parametrizations of microphysical processes.

The observational basis is a dataset of three ground-based vertically pointing Doppler radars (X, Ka and W band) which recorded data during a recent field campaign (TRIPEx, Nov. 2015 - Jan. 2016) at the Jülich Observatory of Cloud Evolution (JOYCE) in Jülich, Germany. The reflectivity differences in the triple-frequency observations have been previously shown to contain information about the average ice particle size and density. The Doppler velocity provides additional constraint to the average particle sedimentation velocity which is a very critical quantity in model parametrizations.

We identified a frontal case during TRIPEx whose temporal and spatial structure is well captured in operational (coarse resolution) model runs. In order to perform sensitivity experiments at different spatial resolutions (down to 100 m hor. res.), we run 200 km wide nested ICON simulations centered over the JOYCE site. In order to compare the model simulations with the radar observations, we use the Passive and Active Microwave Transfer Tool (PAMTRA) which allows to exactly match the ICON assumptions about ice and snow particles (e.g. mass-size relation) using the self-similar Rayleigh-Gans Approximation for their scattering properties.

We run several microphysical experiments to test the influence of microphysical choices like the particle geometry and fall speeds or the assumed size distribution on depositional growth and aggregation and, hence, on the simulated radar reflectivities and Doppler velocities. We further investigate the impact of spatial resolution on the microphysics and overall cloud structure. While the triple-frequency observations provide the strongest constraint to aggregation processes in the lower part of the cloud, the combination of reflectivity and Doppler velocity reveals interesting insights into the early development of the ice particles at cloud top. Preliminary analysis clearly suggests that the major source of model-obs differences are related to the aggregation process rather than to depositional growth.