Real-time radar reflectivity calibration from differential phase measurements

A real-time automated calibration procedure for the reflectivity measurements of the McGill S-band polarimetric operational radar in Montreal has been implemented in the fall of 2012. It is based on measured and theoretical differential phase (Phidp) differences along rain paths of at least 20 km, the theoretical estimate being derived by integrating all the specific differential phase differences Kdp obtained from a suitable Zh - Kdp relationship as suggested by Brandes, (Zh = 5.7 x 10^4 Kdp^1.0746). Information from the RUC2 forecasts provides an initial indication of the height of the 0 C isotherm. Using basic identification techniques based on other polarimetric parameters, care is taken to omit any path with non-liquid precipitation such as a rain/snow mix, hail and graupel. This technique is a variation of the one proposed by Lee and Zawadzki (Journal of Hydrology (2006), p. 83-97), but differs in one important aspect by considering all paths regardless of the intensity of the intervening precipitation, not only those with Phidp differences exceeding 3 deg, thus also allowing its application to cases of light precipitation. In the latter situation, it is not uncommon for measured Phidp differences to be negative on account of significant measurement noise due to our fast scanning (6 rpm) radar antenna. These pairs are not rejected but kept as part of the total Phidp-meas vs Phidp-theor scatter plot. The calibration correction is then derived from the slope of the least-square fit through all the pairs of the measured and of the computed Phidp differences. We show that the inclusion of a sufficiently large number of pairs of Phidp differences inevitably leads to a good estimate of the calibration.