Sizing particles in thick ice clouds using different dual-frequency radar approaches

Description

Measurements from the collocated Ka- and W-band vertically pointing cloud Doppler radars were used to evaluate dual-frequency approaches to retrieve a size parameter of the aggregate particle distributions in thick ice clouds. These approaches are based on non-Rayleigh scattering effects that exist when larger particles are present. The dual-frequency ratio (DFR) and the differential Doppler velocity (DDV) methods were tested using case study observations from the Atmospheric Radiation Measurement (ARM) Southern Great Plains site. Both methods were tuned for retrievals of the exponential distribution slope Λ, which is directly related to the median volume particle size, using appropriate assumptions about ice particle shapes, densities, and fall velocities. Due to measurement errors and uncertainties, meaningful retrievals were generally available for the interval 8 cm-1< Λ < 35 cm-1, although, using the DFR method, slightly larger values of Λ could also be estimated. The expected retrieval errors in the interval between 10 cm-1 and 25 cm-1 were about 30–40% for the DFR-based estimates and about a factor of two larger for the DDV-based estimates. Larger errors for retrievals with the Doppler approach can be explained by higher measurement noise and additional assumptions which are required for this approach. Comparisons of the DDV- and DFR-inferred values of Λ revealed their general consistency with a relative standard deviation between results of both methods being around ~40%, which is within expected retrieval uncertainties. While the DFR approach appears to be more accurate, it requires a 0 dB constraint near cloud tops, which mitigates uncertainties in absolute radar calibrations and differing attenuation paths. The DDV approach generally does not require such a constraint if radar beams are perfectly aligned in vertical (which might not be exactly the case during the SGP observations). Given this, DDV measurements may potentially allow ice particle sizing in situations when DFR constraining is not effective (e.g., in precipitating clouds and in clouds with substantial amounts of supercooled water).