Characterizing falling snow using multifrequency dual-polarization measurements

Combinations of dual‐polarization measurements at various radar frequencies and elevation angles together with realistic scattering computations of snowflakes are compared in order to study how they can be used to characterize falling snow. In the study, we use seven different ice crystal habits, including graupel, and different types of aggregates. The single‐particle scattering is modeled using the discrete‐dipole approximation, and the backscattering cross sections are integrated over exponential size distributions. We take advantage of statistical properties of the large set of scattering results to determine which combination of measurements provide overall best characterization of snow types. We find that a combination of the differential reflectivity and the depolarization ratios, especially in the circular basis, is useful in the characterization, while the dual‐frequency ratios at higher frequencies than the Ku band are almost as good. Also, the assumption of preferential orientation for single ice crystals plays an important role when analyzing the results. These conclusions are tested against both airborne and ground‐based radar measurements obtained during the Global Precipitation Measurement Cold‐season Precipitation Experiment field campaign in 2012. The results show that dual‐polarization and dual‐frequency measurements provide complimentary information and can narrow down the characterization of falling snow considerably. Key Points Dual‐pol and dual‐frequency measurements provide complimentary information Large scattering database can be analyzed by using cluster validation indices Preferential orientation of ice crystals has large influence on characterization