Physical Retrievals from a Multi-frequency Dual-polarization Radar Network during the Mid-latitude Continental Convective Clouds Experiment

This paper presents a retrieval methodology for microphysical retrievals during the Midlatitude Continental Convective Clouds Experiment(MC3E) field project based on a Bayesian inference framework for the ARM Climate Research Facility, and NASA GPM ground radars. The MC3E field campaign was a joint NASA and DOE summer field campaign based in northern Oklahoma during the Summer of 2011. The focus was on multi-scale observations using a wide variety of different instruments, that provided a unique dataset that contains three different radar bands(S, C, and X), as well as other platform observations. Recently there has been an increased interest in multi-radar retrievals, advanced by the CASA (Center for Collaborative Adaptive Sensing of the Atmosphere) system. Most of the past networked retrievals were focused on single frequency radars. MC3E provides a unique opportunity to use field observations to focus on integrating multiple frequencies into the retrieval process. In addition, the dense rain gauge and disdrometer coverage provides us with ample validation opportunities of the derived results, as well as inclusion of a multi-modality probability model into the retrieval process.

The Bayesian retrieval framework described in this paper focuses on retrieving the drop size distributions of clouds in the MC3E campaign, as well as the intrinsic radar product such as reflectivity and specific differential phase. A Bayesian framework allows us to incorporate the different error structures inherent in the large variety of instruments, and represent these in a very natural way within the model. Attenuation correction occurs as a natural effect of this process. This framework can also extend naturally to including non-radar instruments such as disdrometers and profilers, with their associated error matrices.