Due to the advantages of multi-parameter measurements and the technical maturity of radar polarimetry, the national network of weather radars (WSR-88D) is being upgraded with dual-polarization. Meanwhile, phased array technology has attracted much attention in the weather community due to its fast scan capability, possibly lower maintenance costs, and no single-point-of-failure characteristic. Therefore, it is desirable to have both polarimetry and PAR in one radar system for future weather observations. The most significant challenges for a polarimetric PAR (PPAR) is maintaining polarization purity and making unbiased measurements. This is difficult for PPAR in which the two polarizations become non-orthogonal when the beam steers off broadside. It is important to understand this mechanism and to characterize polarimetric phased array antenna performance. This is the goal that is motivating this collaborative research between the Atmospheric Radar Research Center (ARRC) and the IDEAS Laboratory at Purdue University.

Using the Digital Array Radar (DAR), a low-cost, reconfigurable, digital-at-every-element system developed by the IDEAS Laboratory, we created a stable testbed to examine a cylindrical polarimetric array antenna and compare it with a planar array antenna. The DAR is especially apt for this comparison task since it permits the isolation of the antenna array from the backend, thus avoiding the necessity of designing an analog beamformer in each case. Dual-polarization antenna elements are arranged in columns and placed on a 60-degree sector of the cylindrical surface. A horn antenna is used as an ideally-polarized transmit source, and the DAR electronically scans and receives signals in both polarization channels. The received signals are used to form beams both on and off broadside. Dual-pol antenna patterns are characterized and analyzed to assess the performance, which is compared to that of a similar planar array using the transfer matrix compensation for off broadside steering. The issues of capabilities and limitations as well as large-scale implementation will be discussed.