Improvement of polarimetric radar calibration for atmospheric radars

Conventional polarimetric calibration techniques use only measurements of calibration targets located in the boresight direction of the antenna. This calibration has poor performance for atmospheric radars, since weather echoes are spatially distributed and thus the complete antenna pattern contributes to their measurements.

For distributed-target measurements, the boresight polarization isolation of the antenna does not adequately characterize the isolation of the radar polarization channels. This can be explained by the difference between the cross-polar level of the antenna integrated over the complete antenna pattern and the level that is taken at the boresight direction only. Also differences between co-polar and cross-polar angular patterns lead to the fact that observation volumes for these two polarizations are different. This gives unwanted decorrelation between the co-polar and cross-polar reflections of meteorological targets.

In this article we propose a correction method to the point target polarimetric calibration technique. Using this correction technique we are able to take into account the distributed nature of atmospheric targets. This correction is based on light rain measurements with the radar pointed to the zenith. For these measurements rain can be considered as an isotropic target, this assumption is checked by changing the antennas azimuth angle. Furthermore, we calculate the actual antenna polarization isolation level and amount of decorrelation between co- and cross- polar measurements. Based on these calculations correction parameters are introduced into antenna distortion matrices obtained from the conventional polarimetric radar calibration.

In this article we will show that this correction technique improves cross-polar measurements. All measurements in this work are made with the Delft Atmospheric Research Radar, which is an S-band Doppler polarimetric radar.