## Abstract

The sensitivity of polarimetric variables at a 5-cm wavelength to raindrop size and axial ratio is examined using *T*-matrix modeling of the scattering process for gamma raindrop size distributions fitted to tropical rainfall collected at Darwin, Australia. These simulations demonstrate that, while specific differential phase (*K*_{DP})–based estimates of rainfall, attenuation (*A*_{H}), and differential attenuation are less affected by drop size distribution (DSD) variations, large drop occurrence can have significant impacts. Attenuation is sensitive to the occurrence of large drops, which can produce anomalously high values associated with resonance effect scattering. The polarimetric variables are sensitive to the relation between the equivolume diameter and axial ratio. Variations in the assumed form of the raindrop axial ratio can result in significant biases in rainfall and attenuation. Combined rainfall estimators, which include differential reflectivity (*Z*_{DR}), such as *R*(*K*_{DP}, *Z*_{DR}) and *R*(*A*_{H}, *Z*_{DR}) are more robust to both DSD and raindrop axial ratio variations. The results also demonstrate that polarimetric techniques employed to classify the phase of hydrometeors are sensitive to the assumed raindrop axial ratio.

*Corresponding author address:* Dr. T. Keenan, Bureau of Meteorology Research Centre, GPO Box 1289K, Melbourne 3001, Australia.