An Integrated Microphysical Retrieval Framework for Multi-Frequency Dual-Polarization Radar Networks

Multi-Frequency Radar Networks have become an increasingly popular paradigm for both research and quasi-operational setups. CASA (Collaborative Adaptive Sensing of the Atmosphere) have shown the value of radar networks, while the DOE ARM program and the NASA GPM-GV field campaign have started to provide rich multi-frequency network datasets. The polarimetric scattering signature of hydrometeors is highly dependent on the frequency used to measure them, and so multiple frequencies gives us useful information about the particle size distribution (PSD). By leveraging this information, the accuracy of microphysical retrievals can be increased. Little research has been conducted on multi-frequency data for networked radars however.

This paper presents a framework that utilizes multiple radars operating at different frequencies to estimate the drop size distribution. The framework uses a forward variational method to derive a single radar retrieval for each radar, and then the network of instruments is fused into a single network retrieval providing an accurate measure of the drop size distribution field. Multiple examples from both the X-SAPR/C-SAPR radars in the ARM SGP network, and the NASA NPOL/D3R radars in the NASA GPM-GV field campaigns will be shown. When compared to ground disdrometers, we will show NSE error rates as low as 4%, with standardized biases under 10% for both the normalized intercept and the median drop size parameters.

Ground disdrometers, as well as self consistency and measurement reproducibility will be used to show the accuracy of the retrieval. Multiple examples from both the NASA GPM-GV field campaigns such as IFloodS and MC3E, as well as from the ARM Southern Great Plains (SGP) site will be shown. Finally a characterization of the errors will be given.