Monitoring the data quality of the new polarimetric weather radar network of the German Meteorological Service

The German Meteorological Service is introducing new polarimetric weather radar systems in their weather radar network employing the STAR-Mode concept (Simultaneous Transmit and Receive). With the introduction of those systems it is expected that the quality of existing data and products is improved and new products such as a hydrometeor classification can be provided. In order to obtain the full benefit from a polarimetric radar system a comprehensive quality assurance scheme is required. The focus of this scheme is to provide and guarantee high data qualiy, and to monitor the longterm system health in order to guarantee a high availability of the radar system. In this contribution we present how the absolute radar calibration and the calibration of differential reflectivity are monitored.

We have implemented a scheme that is analysing solar interferences based on the operational data, adopting ideas from Holleman et al., 2009. From that analysis we infer the pointing accuracy of the antenna system and subsequently the beamsquint of the antenna assembly, in particular the longterm stability of those antenna parameters. Furthermore, we can monitor the sensitivity of the receiver chain. The monitoring of the receiver chain looks at the absolute and differential sensitivity. From the absolute sensitivity we can characterize parts of the absolute calibration. From the differential sensitivity between the H and V channel we can monitor a component of the ZDR offset (as we do not cover the transmit path). The receiver chain is separately monitored for the two operational pulsewidths 0.8 and 0.4 us. We will show examples of the longterm stability of the receiver chain for various systems, and the ability of the approach to detect quickly alignment issues which can be introduced during e.g. software upgrades. The second element of the monitoring chain employs the so called bird-bath scan. This scan at an elevation of 90° follows every 5 minute volume, again for the two pulse widths. Originally, the main focus of that scan was to monitor the differential phase and power in order to derive the system offsets. Aside from the differential moments this scan is also used to monitor and evaluate the absolute calibration of the radar system (see e.g. Atlas, 2002). This is done in conjunction with an optical disdrometer mounted at the radar site. The measured drop-size distribution of this in-situ instruments provides a direct measure of the radar reflectivity Z . This Z is directly compared with the Z of the first far-field range bin above the radar site. If this range is below the melting layer we find a remarkable agreement between the disdrometer data and the radar data. If we filter for stratiform rain conditions, we can show that the radar system is calibrated within 1 dB, which is the target accuracy for Z. So far, this can be shown for three radar sites. In the future all systems will be equiped with optical disdrometers. The success of this approach is due to the fact that only a limited set of assumptions are needed to make the comparison between the disdrometer and the radar measurement. The results from the birdbath scan are consistent with solar power estimates. At the research radar at Hohenpeißenberg, the results from the analysis has been used to adjust the calibration. A procedure is currently developed to use the result from this monitoring to adjust the calibration for the radar network. It is argued, that this approach is more attractive as it a covers all aspects of the radar equation (especically the radome/antenna part) as compared to the classical calibration, where a well calibrated test-signal generator injects the power directly into the analog receiver. Based on the comparison between birdbath scan and the solar interference data, we can also show that the solar ZDR can be used to monitor and adjust the ZDR offset. This is an important aspect as the birdbath scan requires precipitation at the radar site, which might be sparse depending on the season. We also show that the ZDR computed from the first-lag autocorrelation function is independent of pulse width and temperature effects.