Using polarimetry in mountainous terrain—potentials and challenges

The implementation of polarimetric Doppler radars in mountainous terrain has already been in the planning or implementation phase by several institutions. Within orographically-flat terrain polarimetry has been proven to be a useful addition to radar reflectivity (Z_h) for nowcasting and rainfall rate estimation. The potential benefit of polarimetry in mountainous terrain still needs further examination. Differential reflectivity (Z_dr), correlation coefficient (ρ_hv), and differential propagation phase (φ_dp) tend to be more affected by small contaminations more dominant in mountains compared to radar reflectivity. This study focuses on investigating the influence of contaminations from ground clutter and the melting layer on the accuracy of polarimetric quantities and rainfall estimations. Further aspects to be considered in mountainous terrain include i) the necessity of rapid volume scans with a high spatial resolution since severe weather usually develops and enhances quickly due to orographic forcing and ii) the possibility of using phase instead of reflectivity measurements in areas where the radar beam is shielded. All those aspects will be combined to develop a concept for an optimal usage of polarimetry in mountains.

The sensitivity of polarimetric quantities on ground clutter and melting layer contamination is investigated by using in-phase (I) and quadrature (Q) components of signals measured by the Météo-France C-band polarimetric Doppler radar at Trappes located about 30 km southwest Paris, France. In a first step, I and Q components of various types of ground clutter (i.e., point targets to more complex settings) and various locations within the melting layer are examined. In a second step, I and Q time series from these contaminations are added to I and Q observed within rain. In order to determine the critical level when the superposition of rain and contamination exceeds the measurement accuracy required for polarimetric parameters and rainfall rate, the intensity of the added contamination varies from being 30 dB lower to 30 dB larger than that of rain. This presentation shows the differences in Z_h, Z_dr, ρ_hv, φ_dp, specific differential phase (K_dp) and polarimetric rainfall rates R(Z_h), R(Z_h, Z_dr), R(K_dp), R(K_dp, Z_dr) between the control run using I and Q measurements within rain and the simulated superposition of rain and contamination.

Sensitivity studies of radar reflectivity on ground clutter contamination showed that a measurement accuracy of 1 dB is exceeded when the clutter intensity is ~10 dB larger than Z_h in rain. Polarimetric quantities showed a stronger response to ground clutter contamination. The measurement accuracy was on average exceeded when the ground clutter intensity was ~5 dB larger than the rain intensity. Ground clutter starts to influence considerably the accuracy when the ground clutter contamination is >7 dB for Z_h, >1 dB for Z_dr, and when rain and ground clutter have the same intensity for ρ_hv and φ_dp. As a consequence, rainfall rate estimations using polarimetric quantities tend to be very sensitive to ground clutter contamination compared to R(Z_h). Also, sensitivity studies on temporal and spatial resolution showed high resolution polarimetric quantities tend to be very noisy. In order to maintain the measurement accuracy for polarimetric parameters longer dwell times compared to radar reflectivity are required. This can only be achieved by reducing rotation speed, reducing spatial resolution, or increasing high pulse repetition frequencies.